Add sampled /only: variants (399) for training balance

finetune/configs/sft_v4.yaml

@@ -0,0 +1,38 @@
+# SFT Training Config - v4 with /only: support
+# Usage: accelerate launch --config_file configs/accelerate_multi_gpu.yaml train.py sft --config configs/sft_v4.yaml
+
+model:
+  base: "Qwen/Qwen3-1.7B"
+  output: "qmd-sft-v4"
+
+dataset:
+  name: "data/train_v4"
+  text_field: "text"
+  split: "train"
+  eval_split: 0.1
+
+training:
+  epochs: 3
+  batch_size: 2
+  gradient_accumulation_steps: 4
+  learning_rate: 0.0002
+  max_length: 512
+  warmup_ratio: 0.03
+  lr_scheduler: "cosine"
+
+lora:
+  rank: 16
+  alpha: 32
+  dropout: 0.0
+  target_modules:
+    - "q_proj"
+    - "k_proj"
+    - "v_proj"
+    - "o_proj"
+    - "gate_proj"
+    - "up_proj"
+    - "down_proj"
+
+tracking:
+  project: "qmd-query-expansion"
+  run_name: "sft-1.7B-v4-only-modes"

finetune/data/qmd_only_sampled.jsonl

@@ -0,0 +1,399 @@
+{"input": "what is base jumping? /only:hyde", "output": "hyde: BASE jumping involves parachuting from fixed objects: Buildings, Antennas, Spans (bridges), and Earth (cliffs). Unlike skydiving from aircraft, BASE jumps occur at much lower altitudes, giving jumpers only seconds to deploy their parachute."}
+{"input": "knife skills basics /only:vec", "output": "vec: how to knife skills basics\nvec: complete knife skills basics reference"}
+{"input": "React changelog 2025 /only:vec", "output": "vec: complete React changelog 2025 reference\nvec: how to React changelog 2025"}
+{"input": "cold war key events /only:vec", "output": "vec: understanding cold war key events\nvec: learn about cold war key events"}
+{"input": "how does culture influence ethics /only:vec", "output": "vec: how does culture shape people's ethical beliefs and moral values\nvec: what is the relationship between cultural norms and ethical decision-making"}
+{"input": "what is the difference between ethics and morals /only:lex", "output": "lex: ethics vs morals difference\nlex: ethics morals philosophy distinction\nlex: moral principles ethical systems comparison"}
+{"input": "GitHub recent news November /only:vec", "output": "vec: complete GitHub recent news November reference\nvec: learn about GitHub recent news November"}
+{"input": "how to learn about native american culture /only:vec", "output": "vec: how can I respectfully learn about Native American culture and history\nvec: what are good resources for understanding Indigenous peoples' traditions and heritage"}
+{"input": "decision making psychology /only:vec", "output": "vec: learn about decision making psychology\nvec: how to decision making psychology"}
+{"input": "grilling temperature chart /only:vec", "output": "vec: guide for grilling temperature chart\nvec: understanding grilling temperature chart"}
+{"input": "how do different cultures commemorate death? /only:hyde", "output": "hyde: In Mexico, D\u00eda de los Muertos celebrates deceased loved ones with altars, marigolds, and sugar skulls. Hindu cremation ceremonies release the soul for reincarnation. In Ghana, elaborate fantasy coffins reflect the deceased's life. Japanese Obon festivals welcome ancestral spirits home."}
+{"input": "bronze age collapse /only:vec", "output": "vec: guide for bronze age collapse\nvec: understanding bronze age collapse"}
+{"input": "time zones map /only:vec", "output": "vec: how to time zones map\nvec: complete time zones map reference"}
+{"input": "battery test /only:hyde", "output": "hyde: Test a 12V car battery with a multimeter set to DC volts. A fully charged battery reads 12.6V or higher. Between 12.0-12.4V indicates partial charge. Below 12.0V means the battery is discharged. For a load test, apply a load equal to half the CCA rating for 15 seconds\u2014voltage should stay above 9.6V."}
+{"input": "Han Dynasty China achievements /only:vec", "output": "vec: what were the major achievements and contributions of the han dynasty in china\nvec: why is the han dynasty considered a golden age in chinese history"}
+{"input": "how to brew the perfect cup of tea /only:hyde", "output": "hyde: Water temperature and steep time vary by tea type. Black tea: 200-212\u00b0F for 3-5 minutes. Green tea: 160-180\u00b0F for 2-3 minutes. White tea: 160-185\u00b0F for 4-5 minutes. Oolong: 185-205\u00b0F for 3-5 minutes. Use 1 teaspoon of loose leaf per 8 oz cup. Pre-warm the teapot with hot water for consistent extraction."}
+{"input": "Docker new features 2025 /only:lex", "output": "lex: Docker new features 2025 guide\nlex: Docker new features 2025 best practices\nlex: Docker new features 2025 tutorial"}
+{"input": "weightlifting proper form /only:vec", "output": "vec: guide for weightlifting proper form\nvec: how to weightlifting proper form"}
+{"input": "Byzantine Empire Constantinople /only:vec", "output": "vec: what was the byzantine empire and how long did it last after rome fell\nvec: how did constantinople serve as the capital of the byzantine empire until 1453"}
+{"input": "how does philosophy explore the nature of truth? /only:hyde", "output": "hyde: Philosophy examines truth through several theories. The correspondence theory holds that truth is agreement between a proposition and reality. The coherence theory says a statement is true if it fits consistently within a system of beliefs. The pragmatic theory (James, Dewey) defines truth as what works in practice. Deflationary theories argue that \"true\" adds nothing beyond the assertion itself."}
+{"input": "ancient Egypt pyramids construction /only:vec", "output": "vec: how were the ancient egyptian pyramids constructed without modern technology\nvec: what techniques and labor did ancient egyptians use to build the pyramids at giza"}
+{"input": "what was the enlightenment /only:lex", "output": "lex: Enlightenment 18th century intellectual movement\nlex: Age of Enlightenment reason philosophy\nlex: Enlightenment thinkers Voltaire Locke Kant"}
+{"input": "Vue changelog 2025 /only:vec", "output": "vec: guide for Vue changelog 2025\nvec: understanding Vue changelog 2025"}
+{"input": "TCP vs UDP /only:hyde", "output": "hyde: TCP provides reliable, ordered delivery with acknowledgments and retransmission. UDP is faster but unreliable\u2014packets may arrive out of order or not at all. Use TCP for web, email, file transfer. Use UDP for video streaming, gaming, DNS where speed matters more than reliability."}
+{"input": "how to analyze experimental data /only:hyde", "output": "hyde: Start by cleaning the data: remove outliers using predefined criteria and check for missing values. Calculate descriptive statistics (mean, median, standard deviation). Visualize distributions with histograms or box plots. Apply appropriate statistical tests to evaluate hypotheses. Interpret results in context of your research question and note limitations."}
+{"input": "how to find a reliable realtor /only:vec", "output": "vec: how do you find and vet a trustworthy real estate agent for buying or selling a home\nvec: what qualities and credentials should you look for in a reliable realtor"}
+{"input": "how to build a capsule wardrobe /only:hyde", "output": "hyde: A capsule wardrobe consists of 30-40 versatile pieces that mix and match. Start by choosing a neutral color palette (black, navy, white, beige). Include 2-3 pairs of pants, 5-7 tops, 2 jackets, 2 pairs of shoes, and 1-2 dresses or suits. Remove items you haven't worn in a year. Invest in quality basics over trendy pieces."}
+{"input": "how to sell a car privately? /only:hyde", "output": "hyde: To sell a car privately, first determine a fair price using Kelley Blue Book or Edmunds. Gather the title, maintenance records, and smog certificate. List the car on Craigslist, Facebook Marketplace, or AutoTrader. When meeting buyers, accept cashier's checks or cash. Sign the title over and file a release of liability with your DMV."}
+{"input": "sail boat /only:hyde", "output": "hyde: Sailboats are propelled by wind acting on sails. Common types include dinghies (small, single-hull), keelboats (weighted keel for stability), catamarans (twin hulls), and sloops (single mast, fore-and-aft rigged). Key parts include the hull, mast, boom, jib, mainsail, rudder, and keel."}
+{"input": "grammar punctuation rules /only:lex", "output": "lex: grammar punctuation rules best practices\nlex: grammar punctuation rules documentation\nlex: grammar punctuation rules tutorial"}
+{"input": "tail recursion optimization /only:vec", "output": "vec: what is tail recursion and how does tail call optimization prevent stack overflow\nvec: how do you convert a recursive function to tail recursive form"}
+{"input": "binary search algorithm /only:vec", "output": "vec: how does the binary search algorithm work and what is its time complexity\nvec: how do you implement binary search to find an element in a sorted array"}
+{"input": "how to shoot video in low light /only:hyde", "output": "hyde: For low light video, open your aperture to f/1.4\u2013f/2.8 and lower your shutter speed to 1/50 for 24fps footage. Raise ISO gradually \u2014 modern cameras handle ISO 3200\u20136400 with acceptable noise. Use a fast prime lens and add practical lights in the scene when possible."}
+{"input": "how to boil an egg perfectly /only:hyde", "output": "hyde: Place eggs in a single layer in a pot and cover with cold water by 1 inch. Bring to a rolling boil, then remove from heat and cover. For soft-boiled: 6-7 minutes. For medium: 9-10 minutes. For hard-boiled: 12-13 minutes. Transfer immediately to an ice bath for 5 minutes. Older eggs (7-10 days) peel more easily than fresh ones."}
+{"input": "what is the great wall of china? /only:hyde", "output": "hyde: The Great Wall of China is a series of fortifications built over centuries to protect Chinese states and empires from northern invasions. The most well-known sections were built during the Ming Dynasty (1368-1644). The total length, including all branches and sections across dynasties, is approximately 21,196 kilometers (13,171 miles)."}
+{"input": "stock market basics beginners /only:vec", "output": "vec: guide for stock market basics beginners\nvec: learn about stock market basics beginners"}
+{"input": "how do thought experiments aid philosophical reasoning /only:lex", "output": "lex: thought experiments philosophy reasoning\nlex: philosophical thought experiment trolley problem examples"}
+{"input": "causes of World War I /only:lex", "output": "lex: world war 1 causes\nlex: ww1 assassination archduke franz ferdinand\nlex: causes great war 1914"}
+{"input": "ocean currents patterns /only:lex", "output": "lex: ocean currents patterns tutorial\nlex: ocean currents patterns examples\nlex: ocean currents patterns documentation"}
+{"input": "what are the main sects of islam? /only:lex", "output": "lex: sects of Islam Sunni Shia Sufi\nlex: Islamic denominations branches\nlex: Sunni Shia differences beliefs"}
+{"input": "Kubernetes changelog 2025 /only:lex", "output": "lex: Kubernetes changelog 2025 best practices\nlex: Kubernetes changelog 2025 documentation\nlex: Kubernetes changelog 2025 examples"}
+{"input": "what is the significance of song in worship? /only:lex", "output": "lex: song worship significance religious singing\nlex: worship music congregational singing hymns praise"}
+{"input": "how to create a moon garden? /only:vec", "output": "vec: how to plan and plant a garden designed to be enjoyed at night\nvec: what plants and flowers work best in a moon garden"}
+{"input": "how do religions interpret the concept of sacredness? /only:hyde", "output": "hyde: In Christianity, sacredness is conferred by God's presence\u2014churches, sacraments, and scripture are holy. In Hinduism, sacred rivers like the Ganges and temples house divine energy. Indigenous traditions see sacredness in natural features\u2014mountains, groves, and animals. Islam treats the Quran and Mecca as inviolably sacred."}
+{"input": "what is the significance of the ten commandments /only:hyde", "output": "hyde: The Ten Commandments (Decalogue) were given by God to Moses on Mount Sinai, as recorded in Exodus 20 and Deuteronomy 5. They form the foundational moral code of Judaism and Christianity, covering duties to God (no other gods, no idols, keep the Sabbath) and duties to others (honor parents, do not murder, steal, or lie)."}
+{"input": "how to remove oil stains from clothes /only:lex", "output": "lex: remove oil stains clothing\nlex: grease stain removal fabric\nlex: oil stain laundry treatment"}
+{"input": "how does culture influence identity? /only:hyde", "output": "hyde: Culture shapes identity through language, traditions, values, and social norms internalized from childhood. Family, community, religion, and media all transmit cultural frameworks. Identity is constructed through negotiation between personal experiences and cultural expectations, creating a sense of belonging and self-understanding."}
+{"input": "what is the philosophy of mind /only:lex", "output": "lex: philosophy of mind consciousness mental states\nlex: philosophy of mind problem qualia dualism physicalism"}
+{"input": "how to volunteer for civic initiatives /only:vec", "output": "vec: how can someone find and volunteer for civic engagement and community initiatives\nvec: what are ways to get involved in local civic volunteer opportunities"}
+{"input": "latest findings in climate science /only:lex", "output": "lex: climate science research findings 2025 2026\nlex: climate change latest studies temperature emissions"}
+{"input": "bug fix /only:hyde", "output": "hyde: To fix a bug, first reproduce it reliably and identify the exact conditions that trigger it. Use a debugger or add logging to narrow down the faulty code path. Write a regression test that captures the bug, then modify the code until the test passes."}
+{"input": "where to find heirloom seed suppliers? /only:lex", "output": "lex: heirloom seed suppliers catalog\nlex: buy heirloom seeds online non-GMO\nlex: heirloom vegetable seed company"}
+{"input": "what are smart cities? /only:vec", "output": "vec: what defines a smart city and what technologies do they use\nvec: how do smart cities use IoT sensors and data analytics to improve urban infrastructure"}
+{"input": "what is sacred geometry? /only:hyde", "output": "hyde: Sacred geometry assigns symbolic and spiritual meaning to geometric shapes and proportions found in nature. Key patterns include the Flower of Life (overlapping circles), Metatron's Cube, the golden ratio (1.618), and the Fibonacci spiral. These patterns appear in sunflower seeds, nautilus shells, and ancient temple architecture."}
+{"input": "where to buy raised garden beds? /only:hyde", "output": "hyde: Raised garden beds are available at Home Depot, Lowe's, and garden centers. Online retailers like Gardener's Supply, Amazon, and Birdies offer metal and cedar kits. Cedar is rot-resistant and long-lasting; galvanized steel beds are durable and modern-looking."}
+{"input": "how to start oil painting? /only:lex", "output": "lex: oil painting beginner supplies techniques\nlex: oil painting start canvas brushes paints medium"}
+{"input": "how to do a flip on a trampoline /only:vec", "output": "vec: how do I safely learn to do a backflip on a trampoline\nvec: what is the proper technique for doing flips on a trampoline"}
+{"input": "what is philosophy of mind /only:hyde", "output": "hyde: Philosophy of mind examines the nature of mental states, consciousness, and their relationship to the physical brain. Central questions include the mind-body problem: how do subjective experiences (qualia) arise from neural processes? Key positions include dualism, physicalism, functionalism, and property dualism."}
+{"input": "how to lose weight fast? /only:vec", "output": "vec: what are safe and effective methods to lose weight quickly\nvec: how can I create a calorie deficit to lose weight without harming my health"}
+{"input": "what caused the fall of the roman empire /only:hyde", "output": "hyde: The fall of the Western Roman Empire in 476 AD resulted from multiple factors: military overextension, barbarian invasions (Visigoths, Vandals, Ostrogoths), economic decline from debasement of currency, political instability with rapid emperor turnover, and the shift of power to Constantinople."}
+{"input": "what is the capital of japan /only:hyde", "output": "hyde: Tokyo is the capital city of Japan. It became the capital in 1868 when Emperor Meiji moved the imperial seat from Kyoto. Tokyo, located on the eastern coast of Honshu, is the most populous metropolitan area in the world with over 37 million residents."}
+{"input": "what to pack in a hospital bag for labor? /only:vec", "output": "vec: what items should I pack in my hospital bag before going into labor\nvec: what is a complete packing checklist for the hospital for giving birth"}
+{"input": "what is an anthology? /only:vec", "output": "vec: what is an anthology and how are literary anthologies compiled and organized\nvec: what types of works are typically collected in an anthology such as short stories, poems, or essays"}
+{"input": "what is the role of e-commerce in modern business /only:hyde", "output": "hyde: E-commerce enables businesses to sell products globally without physical storefronts. Companies use platforms like Shopify, Amazon Marketplace, and WooCommerce to reach customers online. In 2024, global e-commerce sales exceeded $6 trillion. Direct-to-consumer (DTC) brands cut out middlemen, while marketplaces aggregate sellers for one-stop shopping."}
+{"input": "how to replace windshield wipers? /only:lex", "output": "lex: replace windshield wipers installation\nlex: change wiper blades car DIY\nlex: windshield wiper replacement size"}
+{"input": "how artificial intelligence is used in healthcare /only:vec", "output": "vec: how is artificial intelligence being applied in healthcare for diagnosis and treatment\nvec: what are the main uses of AI and machine learning in the medical field"}
+{"input": "best techniques for street photography /only:hyde", "output": "hyde: Shoot at f/8 for deep depth of field and zone focus at 3 meters for quick candid shots. Use a 28mm or 35mm lens. Anticipate moments\u2014find good light or backgrounds and wait for subjects to enter the frame. Shoot from the hip to stay inconspicuous."}
+{"input": "what is lean manufacturing /only:hyde", "output": "hyde: Lean manufacturing, derived from the Toyota Production System, aims to minimize waste (muda) while maximizing value. Its five principles: define value from the customer's perspective, map the value stream, create flow, establish pull, and pursue perfection through continuous improvement (kaizen)."}
+{"input": "what is deconstruction /only:hyde", "output": "hyde: Deconstruction, associated with Jacques Derrida, is a method of critical analysis that examines how meaning in texts is constructed through binary oppositions (speech/writing, presence/absence). Derrida argued that meaning is never fixed; it is always deferred through a chain of signifiers. Deconstruction reveals the internal contradictions and assumptions hidden within texts."}
+{"input": "io file /only:vec", "output": "vec: how do you perform file input and output operations in programming languages\nvec: what are the common methods for reading from and writing to files in Python, Java, or C"}
+{"input": "how to choose car speakers? /only:vec", "output": "vec: how do you choose aftermarket car speakers that fit your vehicle and sound preferences\nvec: what is the difference between coaxial and component car speakers and which should you buy"}
+{"input": "how to improve sleep quality /only:lex", "output": "lex: improve sleep quality tips habits\nlex: better sleep hygiene insomnia remedies"}
+{"input": "how do philosophers define happiness /only:lex", "output": "lex: philosophers define happiness philosophy\nlex: happiness eudaimonia Aristotle hedonism\nlex: philosophical theories happiness well-being"}
+{"input": "how to make slime at home /only:vec", "output": "vec: what ingredients and steps do you need to make slime at home\nvec: how to make homemade slime using glue and borax or contact lens solution"}
+{"input": "how to build a writing routine /only:lex", "output": "lex: writing routine daily habit\nlex: build writing practice discipline\nlex: writing schedule productivity"}
+{"input": "how to choose the right camera /only:vec", "output": "vec: how do you choose the right camera for your photography needs and budget?\nvec: what factors should you consider when deciding between DSLR and mirrorless cameras?"}
+{"input": "sail boat /only:lex", "output": "lex: sailboat sailing types rigging\nlex: sailboat buy beginner learn to sail\nlex: sailboat parts hull keel mast"}
+{"input": "Kubernetes ingress controller /only:vec", "output": "vec: what is a kubernetes ingress controller and how does it route external traffic to services\nvec: how do you configure ingress rules for path-based and host-based routing in kubernetes"}
+{"input": "what are the sacred texts of judaism /only:lex", "output": "lex: sacred texts Judaism Torah Talmud\nlex: Jewish scripture Hebrew Bible Tanakh\nlex: Judaism holy books Mishnah"}
+{"input": "how climate change affects farming /only:lex", "output": "lex: climate change agriculture crop yields\nlex: global warming farming drought impact\nlex: climate change food production"}
+{"input": "GraphQL vs REST /only:hyde", "output": "hyde: REST uses fixed endpoints returning predefined data shapes. GraphQL uses one endpoint where clients specify exactly what fields they need, reducing over-fetching. REST is simpler, better cached. GraphQL excels for mobile apps, complex data requirements, and avoiding multiple round trips."}
+{"input": "how to obtain information on federal legislation /only:hyde", "output": "hyde: Congress.gov is the official source for federal legislation. Search by bill number, keyword, or sponsor. Each bill page shows full text, status, cosponsors, committee actions, and vote records. GovTrack.us and ProPublica's Congress API provide additional analysis and tracking tools."}
+{"input": "how to replace car air filter? /only:hyde", "output": "hyde: Open the hood and locate the air filter housing\u2014usually a black plastic box near the engine. Unclip the latches, remove the old filter, and note its orientation. Insert the new filter with the rubber rim facing up, close the housing, and secure the clips. Replace every 12,000-15,000 miles."}
+{"input": "what is a primary election /only:lex", "output": "lex: primary election definition process\nlex: primary election presidential nomination\nlex: open closed primary voting"}
+{"input": "how to support climbing roses? /only:hyde", "output": "hyde: Install a sturdy trellis, arbor, or wire system at least 3 inches from the wall to allow air circulation. Tie canes horizontally with soft plant ties to encourage lateral growth and more blooms. Prune in late winter, removing dead wood and shortening side shoots to 2-3 buds."}
+{"input": "what is a primary election /only:hyde", "output": "hyde: A primary election is a vote held by a political party to choose its candidates for the general election. In a closed primary, only registered party members can vote. In an open primary, any registered voter may participate regardless of party affiliation."}
+{"input": "how does blockchain technology work /only:lex", "output": "lex: blockchain technology distributed ledger\nlex: blockchain cryptography decentralized consensus"}
+{"input": "how to improve car gas mileage? /only:vec", "output": "vec: what are the best ways to improve a car's gas mileage and fuel efficiency\nvec: what driving habits and car maintenance steps help reduce fuel consumption"}
+{"input": "Elasticsearch query DSL /only:lex", "output": "lex: elasticsearch query dsl\nlex: elasticsearch bool must should\nlex: es full text search query"}
+{"input": "what are the themes of to kill a mockingbird? /only:hyde", "output": "hyde: The central themes of To Kill a Mockingbird include racial injustice in the American South, as shown through Tom Robinson's trial. Moral courage is embodied by Atticus Finch, who defends Robinson despite social pressure. The loss of innocence is traced through Scout's growing awareness of prejudice and cruelty in Maycomb, Alabama."}
+{"input": "how to assess a neighborhood safety /only:hyde", "output": "hyde: Check crime maps on sites like CrimeMapping.com or SpotCrime using the ZIP code. Walk the neighborhood at different times of day and night. Look for signs of community investment: maintained properties, street lighting, and active businesses. Talk to residents and visit the local police precinct for crime statistics."}
+{"input": "how to report scientific findings /only:lex", "output": "lex: scientific findings report writing\nlex: research results publication format\nlex: academic paper methodology results"}
+{"input": "how to prepare for a triathlon /only:vec", "output": "vec: what training plan should a beginner follow to prepare for their first triathlon\nvec: how to balance swimming cycling and running workouts when training for a triathlon"}
+{"input": "how to create a moon garden? /only:hyde", "output": "hyde: A moon garden features white and pale-colored flowers, silver foliage, and night-blooming plants that glow under moonlight. Include moonflower (Ipomoea alba), white nicotiana, night-blooming jasmine, dusty miller, and lamb's ear. Add light-colored gravel paths for reflection."}
+{"input": "what is the significance of beauty in philosophy /only:hyde", "output": "hyde: In Plato's Symposium, beauty is a ladder ascending from physical attraction to the Form of Beauty itself. Kant distinguished between the beautiful (harmonious, universal pleasure) and the sublime (overwhelming grandeur). For Hegel, beauty in art reveals truth through sensory form. Contemporary aesthetics debates whether beauty is objective or culturally constructed."}
+{"input": "what is influencer marketing /only:lex", "output": "lex: influencer marketing social media brand promotion\nlex: influencer campaigns Instagram TikTok sponsorship"}
+{"input": "CSS flexbox centering /only:hyde", "output": "hyde: On the container, set display: flex; justify-content: center; align-items: center;. justify-content handles the main axis (horizontal by default), align-items handles the cross axis. Add height: 100vh to center within the viewport. For a single item, margin: auto also works inside flex containers."}
+{"input": "what is the significance of the torah? /only:vec", "output": "vec: what is the Torah and why is it significant in Judaism\nvec: what role does the Torah play in Jewish religious life and law"}
+{"input": "git rebase interactive /only:lex", "output": "lex: git rebase interactive squash\nlex: git rebase -i edit commits\nlex: git squash commits rebase"}
+{"input": "how to hang artwork without nails /only:lex", "output": "lex: hang artwork without nails wall\nlex: picture hanging command strips adhesive hooks"}
+{"input": "fix teeth /only:vec", "output": "vec: what are the options for fixing damaged, chipped, or broken teeth?\nvec: how do dentists repair teeth using crowns, veneers, bonding, and other dental treatments?"}
+{"input": "how to stay motivated daily? /only:lex", "output": "lex: daily motivation habits discipline routine\nlex: stay motivated goals productivity tips"}
+{"input": "how the scientific community addresses research bias /only:lex", "output": "lex: research bias scientific community peer review\nlex: scientific bias mitigation replication reproducibility"}
+{"input": "what is cliffhanger? /only:vec", "output": "vec: what is a cliffhanger in storytelling and how does it create suspense\nvec: how do writers use cliffhangers to keep readers or viewers engaged"}
+{"input": "what is stream of consciousness /only:hyde", "output": "hyde: Stream of consciousness is a narrative technique that presents a character's continuous flow of thoughts, feelings, and sensory impressions as they occur. Pioneered by writers like Virginia Woolf and James Joyce, it mimics the unstructured way the human mind processes experience."}
+{"input": "shell script best practices /only:hyde", "output": "hyde: Start with #!/usr/bin/env bash and set -euo pipefail. Use shellcheck for linting. Quote variables: \"$var\". Use [[ ]] for tests. Handle errors with trap. Use functions for reusability. Avoid parsing ls output\u2014use globs. Prefer printf over echo. Use local variables in functions. Add -- before filenames from user input."}
+{"input": "what is the function of dna /only:lex", "output": "lex: DNA function genetic information\nlex: deoxyribonucleic acid protein synthesis\nlex: DNA replication transcription translation"}
+{"input": "web mail /only:hyde", "output": "hyde: Webmail allows you to access your email through a web browser without installing a desktop client. Popular services include Gmail (mail.google.com), Outlook.com, Yahoo Mail, and ProtonMail. Log in with your credentials to read, compose, and manage messages from any device."}
+{"input": "terraform state management /only:hyde", "output": "hyde: Store state remotely in S3, GCS, or Terraform Cloud\u2014never commit tfstate to git. Configure backend in terraform { backend \"s3\" { bucket = \"my-state\", key = \"prod.tfstate\", region = \"us-east-1\", dynamodb_table = \"tf-locks\" } }. DynamoDB provides state locking to prevent concurrent modifications."}
+{"input": "latest Vue updates /only:vec", "output": "vec: understanding latest Vue updates\nvec: learn about latest Vue updates"}
+{"input": "carbon footprint reduction /only:vec", "output": "vec: guide for carbon footprint reduction\nvec: learn about carbon footprint reduction"}
+{"input": "how does the body maintain homeostasis /only:lex", "output": "lex: homeostasis regulation human body\nlex: negative feedback loop physiology\nlex: body temperature pH blood glucose regulation"}
+{"input": "PostgreSQL indexes explain /only:hyde", "output": "hyde: Run EXPLAIN ANALYZE SELECT... to see the query plan and actual execution time. Look for Seq Scan on large tables\u2014add an index with CREATE INDEX idx_name ON table(column). B-tree indexes work for equality and range queries, GIN for full-text search and arrays, GiST for geometric data."}
+{"input": "what changed in AWS 2025 /only:lex", "output": "lex: what changed in AWS 2025 examples\nlex: what changed in AWS 2025 guide\nlex: what changed in AWS 2025 best practices"}
+{"input": "how do scientists study animal behavior /only:hyde", "output": "hyde: Ethologists use direct observation, video tracking, and GPS telemetry to study animal behavior in natural habitats. Lab experiments control variables to test hypotheses about cognition and social behavior. Focal sampling follows one individual; scan sampling records group behavior at intervals."}
+{"input": "tech fix /only:lex", "output": "lex: tech troubleshooting fix repair computer\nlex: technology fix common problems software hardware\nlex: tech support fix device issue"}
+{"input": "http status codes meaning /only:vec", "output": "vec: what do the common http status codes mean and when should you use each\nvec: how do you choose the right http status code for api responses"}
+{"input": "recent Shopify changes 2025 /only:lex", "output": "lex: recent Shopify changes 2025 examples\nlex: recent Shopify changes 2025 tutorial\nlex: recent Shopify changes 2025 best practices"}
+{"input": "PostgreSQL indexes explain /only:lex", "output": "lex: postgresql index explain analyze\nlex: postgres btree index performance\nlex: create index postgresql"}
+{"input": "http status codes meaning /only:hyde", "output": "hyde: 200 OK success, 201 Created for POST, 204 No Content for DELETE. 400 Bad Request for invalid input, 401 Unauthorized for auth required, 403 Forbidden for insufficient permissions, 404 Not Found. 500 Internal Server Error for unexpected failures, 503 Service Unavailable for temporary issues."}
+{"input": "kindle library /only:lex", "output": "lex: kindle library ebook collection\nlex: Amazon Kindle digital library management\nlex: kindle book organization archive"}
+{"input": "what is the role of family in society /only:lex", "output": "lex: family role society function socialization\nlex: family structure social institution support"}
+{"input": "webhook vs api polling /only:lex", "output": "lex: webhook vs polling api\nlex: push vs pull api pattern\nlex: webhook callback http"}
+{"input": "space exploration changelog 2026 /only:lex", "output": "lex: space exploration changelog 2026 documentation\nlex: space exploration changelog 2026 best practices\nlex: space exploration changelog 2026 guide"}
+{"input": "resilience training programs /only:lex", "output": "lex: resilience training programs mental toughness\nlex: resilience building workplace employee training"}
+{"input": "how to write a scientific research proposal /only:hyde", "output": "hyde: A scientific research proposal typically includes: title, abstract, specific aims, background and significance, preliminary data, research design and methods, timeline, budget and justification, and references. The specific aims page is the most critical \u2014 state the problem, your hypothesis, and 2-3 measurable objectives clearly in one page."}
+{"input": "how to choose farm equipment /only:hyde", "output": "hyde: Match tractor horsepower to your acreage: 25-45 HP for under 50 acres, 45-85 HP for 50-200 acres, and 100+ HP for large operations. Consider PTO power for running implements like mowers and tillers. Evaluate whether two-wheel or four-wheel drive suits your terrain. Used equipment can save 40-60% over new."}
+{"input": "how to enhance creativity? /only:hyde", "output": "hyde: To enhance creativity, practice divergent thinking by generating many ideas without judgment. Keep a daily journal, expose yourself to new experiences, and set aside unstructured time for daydreaming. Research shows that walking, adequate sleep, and constraints can all stimulate creative problem-solving."}
+{"input": "how does culture influence ethics /only:hyde", "output": "hyde: Culture shapes ethics by defining what a society considers right or wrong. Collectivist cultures may prioritize group harmony and duty to family, while individualist cultures emphasize personal autonomy and rights. Cultural relativism argues that moral standards are culturally defined, while universalists hold that some ethical principles transcend culture."}
+{"input": "what are writing prompts? /only:vec", "output": "vec: what are writing prompts and how do writers use them for inspiration\nvec: how do writing prompts help overcome writer's block and spark creativity"}
+{"input": "what is moral philosophy /only:hyde", "output": "hyde: Moral philosophy, or ethics, is the branch of philosophy concerned with questions of right and wrong conduct. It includes three main branches: metaethics (the nature of moral judgments), normative ethics (frameworks like utilitarianism, deontology, and virtue ethics), and applied ethics (specific issues like abortion or euthanasia)."}
+{"input": "how to participate in public policy discussions /only:lex", "output": "lex: participate public policy discussion civic\nlex: public policy engagement town hall\nlex: citizen participation policy advocacy"}
+{"input": "recent React changes 2025 /only:lex", "output": "lex: recent React changes 2025 guide\nlex: recent React changes 2025 best practices\nlex: recent React changes 2025 examples"}
+{"input": "rate limiting algorithms /only:vec", "output": "vec: what algorithms are used for api rate limiting and how do they differ\nvec: how do token bucket and sliding window rate limiting algorithms work"}
+{"input": "how to volunteer for civic initiatives /only:lex", "output": "lex: volunteer civic initiatives community service\nlex: volunteering local government community projects"}
+{"input": "dependency injection benefits /only:lex", "output": "lex: dependency injection di pattern\nlex: di inversion of control ioc\nlex: dependency injection testing"}
+{"input": "what is the significance of algae in ecosystems /only:hyde", "output": "hyde: Algae produce approximately 50% of the world's oxygen through photosynthesis and form the base of aquatic food chains. Phytoplankton, a type of microalgae, supports marine ecosystems by providing energy to zooplankton, fish, and larger organisms."}
+{"input": "json serial /only:lex", "output": "lex: JSON serialization deserialization\nlex: JSON serialize object string\nlex: JSON stringify parse encoding"}
+{"input": "Kubernetes recent news November /only:vec", "output": "vec: how to Kubernetes recent news November\nvec: guide for Kubernetes recent news November"}
+{"input": "how to grow blueberries at home? /only:lex", "output": "lex: grow blueberries home garden\nlex: blueberry bush planting acidic soil\nlex: container blueberry growing care"}
+{"input": "database sharding strategies /only:vec", "output": "vec: what is database sharding and what strategies exist for partitioning data\nvec: how do you choose a shard key and what are the tradeoffs of different sharding approaches"}
+{"input": "craigslist ads /only:vec", "output": "vec: how to post and browse classified ads on Craigslist\nvec: how does Craigslist work for buying, selling, and listing items locally"}
+{"input": "how to protect business data /only:lex", "output": "lex: protect business data security cybersecurity\nlex: data protection encryption backup strategy\nlex: business data security firewall access control"}
+{"input": "what is an elevator pitch /only:vec", "output": "vec: what is an elevator pitch and how do you structure an effective one\nvec: how do you deliver a compelling 30-second pitch for a business idea or job opportunity"}
+{"input": "regex match /only:hyde", "output": "hyde: A regex (regular expression) matches text patterns. Common syntax: `.` matches any character, `*` means zero or more, `+` means one or more, `?` means optional. `[a-z]` matches lowercase letters. `\\d` matches digits. Capture groups use parentheses: `(\\d{3})-(\\d{4})` matches and captures phone number parts. Use `^` for start and `$` for end of line."}
+{"input": "how do i vote in person /only:lex", "output": "lex: vote in person polling place Election Day\nlex: in-person voting process ID requirements"}
+{"input": "what is the significance of logic in philosophy /only:lex", "output": "lex: logic philosophy significance role\nlex: formal logic philosophical argument validity"}
+{"input": "how climate change affects farming /only:vec", "output": "vec: how does rising global temperature affect crop yields and food production\nvec: what effects does climate change have on soil quality and growing seasons for farmers"}
+{"input": "machine learning recent news November /only:vec", "output": "vec: how to machine learning recent news November\nvec: understanding machine learning recent news November"}
+{"input": "kafka consumer groups /only:hyde", "output": "hyde: Consumers with the same group.id share partitions\u2014each partition is consumed by only one consumer in the group. Adding consumers triggers rebalancing. If consumers > partitions, some idle. Offsets track progress per partition. Use enable.auto.commit=false for exactly-once semantics with manual commits."}
+{"input": "what is yoga and its benefits /only:hyde", "output": "hyde: Yoga is an ancient practice combining physical postures (asanas), breathing techniques (pranayama), and meditation. Regular practice improves flexibility, builds strength, reduces stress and anxiety, lowers blood pressure, and enhances sleep quality. Styles range from gentle Hatha to vigorous Vinyasa and Ashtanga."}
+{"input": "how to approach ethical decision-making /only:vec", "output": "vec: what frameworks or steps help with making ethical decisions in difficult situations\nvec: how do you systematically evaluate moral choices when facing an ethical dilemma"}
+{"input": "what are exchange-traded funds (etfs) /only:lex", "output": "lex: exchange-traded funds ETFs investing\nlex: ETF index fund stock market\nlex: ETF vs mutual fund comparison"}
+{"input": "what is the history of the jazz age /only:lex", "output": "lex: Jazz Age history 1920s\nlex: Jazz Age Harlem Renaissance Roaring Twenties\nlex: jazz music history Louis Armstrong"}
+{"input": "what was the impact of the industrial revolution on society? /only:vec", "output": "vec: how did the Industrial Revolution transform society, economy, and daily life?\nvec: what were the major social and economic impacts of the Industrial Revolution on workers and cities?"}
+{"input": "what are the benefits of yoga /only:lex", "output": "lex: yoga benefits health flexibility stress\nlex: yoga physical mental health advantages"}
+{"input": "how to write a business plan /only:lex", "output": "lex: business plan writing template sections\nlex: business plan executive summary financial projections"}
+{"input": "what was the role of the catholic church in the middle ages? /only:lex", "output": "lex: Catholic Church Middle Ages role\nlex: medieval church political power papacy\nlex: Catholic Church feudalism education medieval"}
+{"input": "Vue recent news October /only:lex", "output": "lex: Vue recent news October documentation\nlex: Vue recent news October guide\nlex: Vue recent news October tutorial"}
+{"input": "what is a protagonist? /only:hyde", "output": "hyde: The protagonist is the central character of a narrative, the one whose goals and conflicts drive the plot. The story is told from their perspective or follows their journey. Protagonists are not always heroes\u2014they can be antiheroes or morally ambiguous characters. The antagonist opposes the protagonist, creating the central conflict of the story."}
+{"input": "what is mixed media art? /only:lex", "output": "lex: mixed media art techniques materials\nlex: mixed media collage painting assemblage"}
+{"input": "what are the foundations of feminist ethics /only:hyde", "output": "hyde: Feminist ethics emerged from Carol Gilligan's critique of Kohlberg's moral development theory, arguing that women's moral reasoning emphasizes care and relationships rather than abstract principles of justice. Nel Noddings developed the ethics of care, centering moral life on attentiveness, responsibility, and responsiveness to the needs of particular others."}
+{"input": "renaissance literature /only:lex", "output": "lex: Renaissance literature authors works\nlex: Renaissance literary period Shakespeare Petrarch humanism"}
+{"input": "how to build self-confidence /only:hyde", "output": "hyde: Start by setting small, achievable goals and completing them\u2014each success builds evidence of competence. Practice self-compassion: replace harsh self-criticism with the tone you'd use with a friend. Keep a \"wins\" journal and review it weekly. Gradually expand your comfort zone by doing one slightly uncomfortable thing each day. Confidence grows from accumulated experience, not positive thinking alone."}
+{"input": "how do philosophers define happiness /only:vec", "output": "vec: how have major philosophers throughout history defined happiness and well-being?\nvec: what is the difference between Aristotle's eudaimonia and hedonistic views of happiness?"}
+{"input": "how to manage personal finances /only:lex", "output": "lex: personal finance management\nlex: manage money budgeting saving investing\nlex: personal financial planning"}
+{"input": "yaml vs json config /only:hyde", "output": "hyde: JSON: strict syntax, no comments, explicit quotes, universal parsing. YAML: superset of JSON, allows comments, cleaner for humans, indentation-based. Use JSON for data interchange, APIs, when strict parsing needed. Use YAML for configs (Docker Compose, Kubernetes, CI/CD) where human editing is common. YAML gotchas: Norway problem (NO parsed as false), inconsistent indentation."}
+{"input": "Kubernetes changelog 2026 /only:lex", "output": "lex: Kubernetes changelog 2026 documentation\nlex: Kubernetes changelog 2026 examples\nlex: Kubernetes changelog 2026 tutorial"}
+{"input": "retirement planning strategies /only:lex", "output": "lex: retirement planning strategies guide\nlex: retirement planning strategies documentation\nlex: retirement planning strategies examples"}
+{"input": "how to participate in public policy discussions /only:vec", "output": "vec: how can citizens effectively participate in public policy discussions and influence government decisions?\nvec: what are the ways individuals can engage in public policy debates at the local, state, and federal level?"}
+{"input": "how to repair a leaky faucet /only:vec", "output": "vec: how to fix a dripping faucet by replacing the washer or cartridge\nvec: what are the step-by-step instructions for repairing a leaky kitchen or bathroom faucet"}
+{"input": "Meiji Restoration Japan /only:hyde", "output": "hyde: The Meiji Restoration (1868) ended 250 years of Tokugawa shogunate rule, restoring imperial power under Emperor Meiji. Japan rapidly industrialized and westernized: abolished feudalism, created national army, built railways, established constitution (1889). Slogan: 'Rich country, strong army.' Japan defeated China (1895) and Russia (1905), becoming a world power within 50 years."}
+{"input": "AI recent news December /only:lex", "output": "lex: AI recent news December documentation\nlex: AI recent news December guide\nlex: AI recent news December best practices"}
+{"input": "ocean plastic pollution /only:lex", "output": "lex: ocean plastic pollution examples\nlex: ocean plastic pollution guide\nlex: ocean plastic pollution documentation"}
+{"input": "React context vs Redux /only:hyde", "output": "hyde: Context is built-in, simple for low-frequency updates like themes and auth. Redux adds boilerplate but provides devtools, middleware, time-travel debugging, predictable updates. Context re-renders all consumers on any change; Redux allows granular subscriptions. Use Context for simple cases, Redux for complex state logic."}
+{"input": "IPv4 vs IPv6 /only:lex", "output": "lex: ipv4 ipv6 difference\nlex: ipv6 address format\nlex: ipv4 exhaustion ipv6 transition"}
+{"input": "how to repair a leaky faucet /only:hyde", "output": "hyde: Turn off the water supply valves under the sink. Remove the faucet handle by unscrewing the decorative cap and handle screw. Pull out the stem or cartridge. For compression faucets, replace the rubber washer and O-ring. For cartridge faucets, replace the entire cartridge. Reassemble, turn the water back on, and test for leaks."}
+{"input": "what is the role of faith in spirituality /only:hyde", "output": "hyde: Faith in spirituality serves as the foundation for trust in a reality beyond the material world. It enables surrender to uncertainty and provides a framework for interpreting suffering and purpose. Unlike dogmatic belief, spiritual faith often involves personal experience\u2014a felt sense of connection to something greater that sustains practice through doubt and difficulty."}
+{"input": "what causes market volatility /only:vec", "output": "vec: what economic and geopolitical factors cause stock market volatility\nvec: why do financial markets experience sudden price swings and instability"}
+{"input": "calculus derivatives explained /only:lex", "output": "lex: calculus derivatives explained best practices\nlex: calculus derivatives explained examples\nlex: calculus derivatives explained documentation"}
+{"input": "machine learning changelog 2025 /only:lex", "output": "lex: machine learning changelog 2025 examples\nlex: machine learning changelog 2025 guide\nlex: machine learning changelog 2025 best practices"}
+{"input": "how to kayak for the first time /only:vec", "output": "vec: what should a beginner know before going kayaking for the first time\nvec: how do I paddle and balance a kayak as a first-time kayaker"}
+{"input": "how to participate in earth hour? /only:lex", "output": "lex: Earth Hour participation lights off event\nlex: Earth Hour date 2026 how to join"}
+{"input": "TypeScript changelog 2025 /only:lex", "output": "lex: TypeScript changelog 2025 documentation\nlex: TypeScript changelog 2025 examples\nlex: TypeScript changelog 2025 guide"}
+{"input": "what is the importance of spiritual leadership? /only:vec", "output": "vec: how does spiritual leadership influence organizations and their members\nvec: what role does spiritual leadership play in providing meaning and purpose at work"}
+{"input": "rivers that cross multiple countries /only:lex", "output": "lex: rivers that cross multiple countries documentation\nlex: rivers that cross multiple countries tutorial\nlex: rivers that cross multiple countries guide"}
+{"input": "async web /only:vec", "output": "vec: how do asynchronous programming patterns work in web development and API requests?\nvec: what are the best async web frameworks for building non-blocking HTTP servers?"}
+{"input": "where to find eco-friendly furniture /only:vec", "output": "vec: where can I buy eco-friendly and sustainably made furniture\nvec: what brands and stores sell furniture made from sustainable or recycled materials"}
+{"input": "when to replace windshield wipers? /only:vec", "output": "vec: how often should windshield wipers be replaced and what are signs they need changing\nvec: what are the signs that windshield wiper blades are worn out and need replacement"}
+{"input": "what are the characteristics of a just society /only:lex", "output": "lex: just society characteristics principles fairness\nlex: social justice equality Rawls distributive justice"}
+{"input": "how to manage sibling rivalry? /only:hyde", "output": "hyde: Avoid comparing siblings to each other. Give each child individual attention and acknowledge their unique strengths. Teach conflict resolution skills rather than always intervening. Set clear family rules about respectful behavior and let children solve minor disputes themselves."}
+{"input": "how to build self-confidence /only:vec", "output": "vec: what are practical strategies for building self-confidence and overcoming self-doubt\nvec: how can someone develop greater self-confidence through daily habits and mindset shifts"}
+{"input": "React changelog 2026 /only:lex", "output": "lex: React changelog 2026 tutorial\nlex: React changelog 2026 examples\nlex: React changelog 2026 documentation"}
+{"input": "what are the best soil types for roses /only:vec", "output": "vec: what type of soil do roses grow best in and how should it be prepared\nvec: what soil pH and composition are ideal for growing healthy rose bushes"}
+{"input": "how does compound interest work /only:hyde", "output": "hyde: Compound interest is calculated on both the principal and accumulated interest. The formula is A = P(1 + r/n)^(nt), where P is principal, r is annual rate, n is compounding frequency, and t is time in years. Monthly compounding on $10,000 at 5% yields $16,470 after 10 years."}
+{"input": "tail recursion optimization /only:hyde", "output": "hyde: Tail recursion: recursive call is the last operation, no work after it returns. TCO reuses stack frame instead of adding new one\u2014prevents stack overflow. Convert by passing accumulated result as parameter: factorial(n, acc=1) { return n <= 1 ? acc : factorial(n-1, n*acc); }. Not all languages implement TCO\u2014JavaScript in strict mode, Scheme yes, Python no."}
+{"input": "TypeScript generics /only:hyde", "output": "hyde: Generics let you write flexible, reusable code while maintaining type safety. Declare with angle brackets: function identity<T>(arg: T): T { return arg; }. Add constraints with extends: function getLength<T extends { length: number }>(item: T): number { return item.length; }."}
+{"input": "what is the importance of spiritual leadership? /only:hyde", "output": "hyde: Spiritual leadership theory proposes that leaders who foster a sense of calling, meaning, and membership create more engaged and productive organizations. It emphasizes vision, altruistic love, and hope as core values that transcend traditional management."}
+{"input": "how does culture influence identity? /only:vec", "output": "vec: how does culture shape a person's sense of identity\nvec: in what ways do cultural values and traditions influence who we become"}
+{"input": "digital transformation strategies /only:vec", "output": "vec: what strategies do organizations use to drive successful digital transformation\nvec: how do enterprises plan and execute a digital transformation initiative"}
+{"input": "how to build confidence in social situations? /only:hyde", "output": "hyde: Start small: make eye contact and greet one new person at each event. Prepare a few open-ended questions in advance. Focus on listening rather than performing. After each interaction, note what went well. Gradual exposure reduces anxiety over time\u2014the more you practice, the more natural conversations become."}
+{"input": "how to analyze a political candidate's stance /only:lex", "output": "lex: analyze political candidate stance positions\nlex: candidate policy positions voting record\nlex: compare political candidates issues"}
+{"input": "JWT token refresh /only:vec", "output": "vec: how does the jwt refresh token flow work for maintaining user sessions\nvec: what is the difference between access tokens and refresh tokens in jwt authentication"}
+{"input": "where to buy used cars online /only:lex", "output": "lex: buy used cars online marketplace\nlex: certified pre-owned cars website\nlex: online used car dealers Carvana AutoTrader"}
+{"input": "GitHub Actions workflow /only:lex", "output": "lex: GitHub Actions workflow guide\nlex: GitHub Actions workflow examples\nlex: GitHub Actions workflow documentation"}
+{"input": "git push /only:hyde", "output": "hyde: Use `git push origin main` to push your local main branch to the remote. For a new branch, use `git push -u origin feature-branch` to set the upstream tracking reference. If the push is rejected because the remote has new commits, run `git pull --rebase` first, then push again."}
+{"input": "what is the categorical imperative /only:vec", "output": "vec: what is Kant's categorical imperative and how does it function as a moral principle\nvec: how does the categorical imperative test whether an action is morally permissible"}
+{"input": "how to build a personal brand /only:hyde", "output": "hyde: Build your personal brand by defining your niche and unique value proposition. Create consistent profiles across LinkedIn, Twitter, and a personal website. Publish content regularly\u2014blog posts, videos, or podcasts\u2014that demonstrates your expertise. Engage authentically with your audience and network at industry events."}
+{"input": "how augmented reality is applied in different fields /only:hyde", "output": "hyde: Augmented reality overlays digital content onto the real world and is applied across many fields. In healthcare, surgeons use AR to visualize anatomy during procedures. In education, AR apps bring textbook content to life in 3D. Retailers like IKEA use AR to let customers preview furniture in their homes. In manufacturing, AR guides workers through assembly with step-by-step overlays."}
+{"input": "memory leak debugging /only:lex", "output": "lex: memory leak debug profiler\nlex: memory leak detection tools\nlex: heap dump memory analysis"}
+{"input": "what is zero waste? /only:hyde", "output": "hyde: Zero waste is a philosophy and lifestyle aiming to send nothing to landfills by reducing consumption, reusing items, recycling, and composting. Practical steps include using reusable bags, bottles, and containers, buying in bulk, composting food scraps, and choosing products with minimal or recyclable packaging."}
+{"input": "how to enhance concentration /only:hyde", "output": "hyde: Improve concentration by eliminating distractions: silence notifications, use website blockers, and work in a quiet environment. The Pomodoro Technique\u201425 minutes of focused work followed by a 5-minute break\u2014builds sustained attention. Regular exercise, adequate sleep (7-9 hours), and mindfulness meditation physically strengthen the brain's prefrontal cortex."}
+{"input": "how to improve workplace productivity /only:vec", "output": "vec: what strategies and techniques can improve productivity in the workplace\nvec: how can employees and managers increase work output and reduce wasted time"}
+{"input": "how to fix car key fob? /only:vec", "output": "vec: how do I fix a car key fob that stopped working\nvec: how to replace the battery or reprogram a car key fob"}
+{"input": "what are the key features of taoist philosophy? /only:vec", "output": "vec: what are the central concepts and key features of Taoist philosophy?\nvec: how does Taoism emphasize living in harmony with the Tao and the concept of wu wei?"}
+{"input": "how to write a standout personal statement /only:hyde", "output": "hyde: Open with a vivid, specific anecdote\u2014not a generic quote. Show rather than tell by describing experiences that shaped your goals. Connect your past to your intended field of study. Be authentic; admissions officers read thousands of essays and recognize genuine voice immediately."}
+{"input": "what is outdoor survival training? /only:hyde", "output": "hyde: Outdoor survival training teaches skills needed to stay alive in wilderness emergencies. Core topics include building emergency shelters from natural materials, finding and purifying water, starting fire without matches using a ferro rod or bow drill, signaling for rescue, and basic navigation without GPS. Courses range from weekend workshops to multi-week immersive programs."}
+{"input": "how do philosophers approach the meaning of life /only:hyde", "output": "hyde: Existentialists like Sartre argued life has no inherent meaning\u2014we must create it through our choices. Aristotle proposed eudaimonia (flourishing) as life's purpose. Camus explored the absurd, suggesting we must find meaning despite an indifferent universe. Eastern philosophy often points to liberation from suffering."}
+{"input": "japanese hiragana katakana /only:vec", "output": "vec: complete japanese hiragana katakana reference\nvec: guide for japanese hiragana katakana"}
+{"input": "what changed in machine learning 2025 /only:lex", "output": "lex: what changed in machine learning 2025 guide\nlex: what changed in machine learning 2025 best practices\nlex: what changed in machine learning 2025 documentation"}
+{"input": "how to create a home office space /only:hyde", "output": "hyde: Set up your home office in a quiet room with natural light. Invest in an ergonomic chair with lumbar support and a desk at elbow height (28-30 inches). Position your monitor at arm's length with the top at eye level. Use a desk lamp with 4000-5000K color temperature. Keep cables organized and add a plant\u2014studies show greenery reduces stress and improves focus."}
+{"input": "latest Shopify updates /only:vec", "output": "vec: complete latest Shopify updates reference\nvec: guide for latest Shopify updates"}
+{"input": "Next.js new features 2026 /only:vec", "output": "vec: learn about Next.js new features 2026\nvec: how to Next.js new features 2026"}
+{"input": "how did the roman empire impact culture? /only:hyde", "output": "hyde: The Roman Empire's cultural legacy includes Latin (the root of Romance languages), Roman law (the basis of civil law systems worldwide), architectural innovations like arches, aqueducts, and concrete, republican government concepts, road networks, and the spread of Christianity. Roman art, literature, and engineering influenced Western civilization for centuries."}
+{"input": "how to participate in a town hall meeting /only:hyde", "output": "hyde: Check your local government website or social media for upcoming town hall schedules. Arrive early and sign up to speak if required. Prepare a concise statement (usually 2-3 minutes). Stay respectful and on-topic. Bring supporting data or personal stories to strengthen your point."}
+{"input": "how to be a good listener /only:vec", "output": "vec: how can I become a better and more active listener in conversations\nvec: what techniques improve listening skills and show empathy"}
+{"input": "web socket /only:hyde", "output": "hyde: WebSocket provides full-duplex communication over a single TCP connection. After an HTTP upgrade handshake, client and server can send messages in both directions without polling. Use `new WebSocket('ws://host/path')` on the client and a library like ws on the server."}
+{"input": "apache kafka partitions /only:hyde", "output": "hyde: Partitions enable parallelism\u2014each partition is consumed by one consumer in a group. Messages with same key go to same partition, preserving order per key. More partitions = more throughput but more overhead. Start with partitions = max(expected throughput / partition throughput, consumer count). Can't reduce partitions, only increase."}
+{"input": "inflation effects on savings /only:vec", "output": "vec: guide for inflation effects on savings\nvec: complete inflation effects on savings reference"}
+{"input": "nginx location block /only:vec", "output": "vec: how do nginx location blocks work and in what order are they matched\nvec: what is the syntax for nginx location directives including prefix and regex matching"}
+{"input": "IPv4 vs IPv6 /only:vec", "output": "vec: what are the key differences between ipv4 and ipv6 addressing\nvec: why is ipv6 necessary and how does the transition from ipv4 work"}
+{"input": "how to raise startup capital /only:hyde", "output": "hyde: Startup capital can come from bootstrapping, friends and family, angel investors, venture capital firms, crowdfunding platforms like Kickstarter, or government grants. Prepare a pitch deck with your business model, market size, traction metrics, and financial projections before approaching investors."}
+{"input": "tectonic plate boundaries /only:lex", "output": "lex: tectonic plate boundaries examples\nlex: tectonic plate boundaries documentation\nlex: tectonic plate boundaries best practices"}
+{"input": "surfing wave types /only:lex", "output": "lex: surfing wave types tutorial\nlex: surfing wave types examples\nlex: surfing wave types documentation"}
+{"input": "how to use a light meter /only:lex", "output": "lex: light meter photography exposure reading\nlex: incident reflected light meter settings"}
+{"input": "what is literary parody? /only:hyde", "output": "hyde: Literary parody imitates the style, conventions, or content of a specific work or genre for comedic or critical effect. It exaggerates distinctive features to expose flaws or absurdities. Examples include Don Quixote (parodying chivalric romances), Northanger Abbey (Gothic novels), and The Hitchhiker's Guide to the Galaxy (science fiction tropes)."}
+{"input": "how to install peel and stick wallpaper /only:lex", "output": "lex: peel and stick wallpaper installation\nlex: self-adhesive wallpaper apply walls"}
+{"input": "how to meditate for beginners /only:vec", "output": "vec: how do beginners start a daily meditation practice from scratch\nvec: what are simple meditation techniques for people who have never meditated before"}
+{"input": "renaissance sculpture techniques /only:lex", "output": "lex: renaissance sculpture techniques documentation\nlex: renaissance sculpture techniques examples\nlex: renaissance sculpture techniques best practices"}
+{"input": "how to save money effectively /only:vec", "output": "vec: what are effective strategies and habits for saving money consistently\nvec: how can I create a budget and save more money each month"}
+{"input": "GraphQL subscriptions websocket /only:vec", "output": "vec: how do graphql subscriptions work for real-time data updates\nvec: what is the underlying protocol for graphql subscriptions and how do you implement them"}
+{"input": "what is the significance of algae in ecosystems /only:lex", "output": "lex: algae ecosystem role food chain\nlex: algae oxygen production aquatic ecosystems\nlex: algae photosynthesis carbon cycle"}
+{"input": "React recent news November /only:lex", "output": "lex: React recent news November best practices\nlex: React recent news November examples\nlex: React recent news November guide"}
+{"input": "sailing adventures /only:lex", "output": "lex: sailing adventure trips voyages\nlex: sailing vacation destinations cruises\nlex: ocean sailing expedition"}
+{"input": "Elasticsearch query DSL /only:lex", "output": "lex: elasticsearch query dsl\nlex: elasticsearch bool must should\nlex: es full text search query"}
+{"input": "what are the elements of short stories? /only:hyde", "output": "hyde: The essential elements of a short story are plot (the sequence of events), character (the people involved), setting (time and place), conflict (the central struggle), theme (the underlying message), and point of view (the narrative perspective). Short stories typically focus on a single incident."}
+{"input": "Docker changelog 2026 /only:vec", "output": "vec: understanding Docker changelog 2026\nvec: complete Docker changelog 2026 reference"}
+{"input": "what is the role of sacred music in worship? /only:vec", "output": "vec: what role does sacred music play in religious worship services across different faiths\nvec: how do hymns, chants, and liturgical music enhance the experience of communal worship"}
+{"input": "hair cut /only:vec", "output": "vec: what are the popular haircut styles and how to choose the right one\nvec: how to communicate what haircut you want to a stylist or barber"}
+{"input": "Han Dynasty China achievements /only:lex", "output": "lex: han dynasty china achievements\nlex: han dynasty 206 bc history\nlex: ancient china han empire"}
+{"input": "how to increase productivity at work? /only:lex", "output": "lex: productivity work increase tips\nlex: workplace productivity time management techniques"}
+{"input": "what is the significance of the alhambra? /only:vec", "output": "vec: why is the Alhambra in Granada, Spain considered a masterpiece of Islamic architecture?\nvec: what is the cultural and historical significance of the Alhambra palace?"}
+{"input": "what is bioethics /only:lex", "output": "lex: bioethics definition medical ethics biology\nlex: bioethics issues euthanasia cloning genetic engineering"}
+{"input": "how to build a successful brand /only:lex", "output": "lex: brand building strategy identity positioning\nlex: brand identity logo messaging target audience"}
+{"input": "stellar cartography /only:vec", "output": "vec: what is stellar cartography and how do astronomers map the positions and movements of stars?\nvec: what tools and surveys are used to create detailed maps of stars in the galaxy?"}
+{"input": "what are the key principles of confucianism? /only:hyde", "output": "hyde: The key principles of Confucianism include Ren (benevolence/humaneness), Li (ritual propriety), Xiao (filial piety), Yi (righteousness), and Zhi (wisdom). The Five Relationships define social bonds: ruler-subject, parent-child, husband-wife, elder-younger sibling, and friend-friend. Each relationship carries reciprocal obligations."}
+{"input": "what changed in GitHub 2026 /only:lex", "output": "lex: what changed in GitHub 2026 tutorial\nlex: what changed in GitHub 2026 guide\nlex: what changed in GitHub 2026 documentation"}
+{"input": "when to introduce solid foods to a baby? /only:hyde", "output": "hyde: Most pediatricians recommend introducing solid foods around 6 months of age. Signs of readiness include sitting up with support, showing interest in food, and loss of the tongue-thrust reflex. Start with single-ingredient purees like sweet potato, avocado, or iron-fortified cereal, one new food every 3-5 days."}
+{"input": "how to mix modern and vintage decor /only:vec", "output": "vec: how do you blend vintage furniture and antique pieces with modern interior design elements\nvec: what are effective ways to combine mid-century or antique decor with contemporary minimalist style"}
+{"input": "latest climate tech updates /only:vec", "output": "vec: understanding latest climate tech updates\nvec: complete latest climate tech updates reference"}
+{"input": "SQL injection prevention /only:vec", "output": "vec: how to prevent sql injection attacks in web applications\nvec: why are parameterized queries and prepared statements important for database security"}
+{"input": "how to use a ring light /only:hyde", "output": "hyde: Place the ring light directly in front of your face at eye level, with the camera positioned in the center of the ring. Keep the light 12-24 inches from your face for an even, shadow-free glow. Adjust brightness to avoid overexposure. The circular catchlights in the eyes are a signature look."}
+{"input": "how to increase home resale value /only:lex", "output": "lex: increase home resale value renovations\nlex: home improvement ROI property value"}
+{"input": "how to set financial goals /only:lex", "output": "lex: set financial goals planning budget\nlex: financial goal setting SMART savings\nlex: personal finance goals short long term"}
+{"input": "GitHub new features 2025 /only:lex", "output": "lex: GitHub new features 2025 guide\nlex: GitHub new features 2025 tutorial\nlex: GitHub new features 2025 examples"}
+{"input": "how to handle inflation impact /only:vec", "output": "vec: how can individuals protect their finances and manage the impact of high inflation\nvec: what financial strategies help people cope with rising prices and reduced purchasing power"}
+{"input": "Python new features 2026 /only:vec", "output": "vec: guide for Python new features 2026\nvec: complete Python new features 2026 reference"}
+{"input": "how to manage anxiety naturally /only:hyde", "output": "hyde: Natural anxiety management includes regular aerobic exercise (30 minutes, 5 days a week), diaphragmatic breathing, progressive muscle relaxation, and limiting caffeine and alcohol. Cognitive behavioral techniques like thought journaling help identify and challenge anxious thinking patterns. Herbal supplements such as chamomile and ashwagandha show some evidence of benefit."}
+{"input": "how to talk to kids about bullying? /only:lex", "output": "lex: talk children bullying conversation advice\nlex: kids bullying prevention parent discussion"}
+{"input": "what is impact investing? /only:hyde", "output": "hyde: Impact investing directs capital toward companies and projects that generate measurable social or environmental benefits alongside financial returns. Unlike ESG screening, which excludes harmful sectors, impact investing actively targets positive outcomes \u2014 such as affordable housing, renewable energy, or microfinance. The Global Impact Investing Network (GIIN) estimates the market at over $1 trillion."}
+{"input": "what is competitive analysis /only:vec", "output": "vec: what is competitive analysis in business and how do companies use it to inform strategy\nvec: what frameworks and methods are used to conduct a competitive analysis of rival companies"}
+{"input": "latest updates on the ukraine conflict /only:vec", "output": "vec: what are the most recent developments in the Russia-Ukraine war as of 2025-2026?\nvec: what is the current status of the Ukraine conflict including ceasefire talks and territorial changes?"}
+{"input": "what is the role of research institutions /only:lex", "output": "lex: research institutions universities role science\nlex: research institutions funding labs innovation"}
+{"input": "how to enhance positive social impact? /only:lex", "output": "lex: enhance social impact community\nlex: positive social impact strategies nonprofit\nlex: social change community engagement"}
+{"input": "what is the significance of pilgrimage in religion? /only:hyde", "output": "hyde: Pilgrimage holds deep significance across religions. Muslims perform Hajj to Mecca as one of the Five Pillars. Christians journey to Jerusalem, Rome, and Santiago de Compostela. Hindus bathe in the Ganges at Varanasi. The physical journey symbolizes an inner spiritual transformation\u2014leaving ordinary life, enduring hardship, and arriving at a sacred place of renewal and encounter with the divine."}
+{"input": "how to stay updated on global affairs /only:lex", "output": "lex: global affairs news sources current events\nlex: world news reliable sources daily updates"}
+{"input": "where to buy affordable art prints /only:hyde", "output": "hyde: Affordable art prints are available on Society6, Redbubble, and Etsy, where independent artists sell prints starting at $15\u2013$30. IKEA offers framed prints under $20. For museum-quality reproductions, check Artsy or Saatchi Art's prints section. King & McGaw specializes in licensed fine art reproductions at mid-range prices."}
+{"input": "building resilience /only:hyde", "output": "hyde: Building resilience involves developing a growth mindset, maintaining social connections, and practicing self-care. Reframe setbacks as learning opportunities. Cultivate problem-solving skills rather than ruminating on what went wrong. Regular exercise, adequate sleep, and mindfulness strengthen your capacity to recover from stress. Resilient people accept what they cannot control and focus energy on what they can."}
+{"input": "what is the significance of community in spirituality? /only:vec", "output": "vec: why is community considered important in spiritual and religious practice?\nvec: how does belonging to a spiritual community enhance personal faith and practice?"}
+{"input": "what is the significance of the hajj in islam? /only:lex", "output": "lex: Hajj Islam pilgrimage Mecca significance\nlex: Hajj pillar Islam Kaaba rituals"}
+{"input": "what is the significance of community in ethics /only:vec", "output": "vec: what role does community play in ethical theory and moral life\nvec: how does communitarian philosophy view the relationship between community and ethics"}
+{"input": "how to reduce stress /only:hyde", "output": "hyde: Regular physical activity releases endorphins that naturally reduce stress. Practice deep breathing: inhale for 4 counts, hold for 4, exhale for 6. Other effective strategies include progressive muscle relaxation, journaling, limiting caffeine, and maintaining a consistent sleep schedule of 7-9 hours."}
+{"input": "how to improve sleep quality /only:hyde", "output": "hyde: Maintain a consistent sleep schedule, even on weekends. Keep your bedroom cool (65-68\u00b0F), dark, and quiet. Avoid screens for 30 minutes before bed. Limit caffeine after noon. Regular exercise improves sleep, but finish workouts at least 3 hours before bedtime."}
+{"input": "ancient Greece democracy Athens /only:lex", "output": "lex: ancient greece democracy athens\nlex: athenian democracy 5th century bc\nlex: greek democracy origins"}
+{"input": "solid principles explained /only:hyde", "output": "hyde: SOLID: Single Responsibility (one reason to change), Open/Closed (open for extension, closed for modification), Liskov Substitution (subtypes substitutable for base types), Interface Segregation (many specific interfaces over one general), Dependency Inversion (depend on abstractions not concretions). Following SOLID produces loosely coupled, testable, maintainable code."}
+{"input": "TypeScript changelog 2026 /only:vec", "output": "vec: understanding TypeScript changelog 2026\nvec: how to TypeScript changelog 2026"}
+{"input": "mac address vs ip address /only:vec", "output": "vec: what is the difference between a mac address and an ip address in networking\nvec: how do mac addresses and ip addresses work together for network communication"}
+{"input": "protobuf vs json /only:hyde", "output": "hyde: JSON: human-readable, self-describing, universal support, larger payload. Protobuf: binary format, 3-10x smaller, faster serialization, requires schema (.proto files), strong typing. Use JSON for public APIs, debugging, human interaction. Use Protobuf for internal microservices, high-throughput systems, gRPC. Schema evolution with field numbers enables backward compatibility."}
+{"input": "how to find emotional support /only:hyde", "output": "hyde: Find emotional support through multiple channels: talk to a trusted friend or family member. Contact a therapist through Psychology Today's directory or your insurance provider. Call the 988 Suicide and Crisis Lifeline (dial 988) for immediate help. Join support groups through NAMI or local community centers. Online therapy platforms like BetterHelp and Talkspace offer accessible counseling."}
+{"input": "how to plant a wildflower meadow? /only:vec", "output": "vec: how do I plant and establish a wildflower meadow in my yard\nvec: what steps are needed to create a wildflower meadow from seed"}
+{"input": "how to succeed in a digital marketing career? /only:vec", "output": "vec: what skills and experience do you need to build a successful digital marketing career\nvec: how to get started in digital marketing and advance to senior roles"}
+{"input": "what is the difference between realism and idealism /only:lex", "output": "lex: realism idealism philosophy difference\nlex: realism vs idealism metaphysics epistemology\nlex: philosophical realism idealism comparison"}
+{"input": "ai-driven analytics /only:hyde", "output": "hyde: AI-driven analytics uses machine learning algorithms to automatically detect patterns, anomalies, and trends in large datasets. Unlike traditional BI tools, AI analytics can generate predictive forecasts, perform natural language queries, and surface insights without manual configuration."}
+{"input": "how does the social contract theory explain governance /only:vec", "output": "vec: how does social contract theory explain the legitimacy of government\nvec: what did Hobbes, Locke, and Rousseau argue about the social contract and governance"}
+{"input": "how to grow tomatoes at home? /only:lex", "output": "lex: grow tomatoes home garden\nlex: tomato plant care watering sunlight\nlex: container tomatoes growing tips"}
+{"input": "what is the significance of archetypes? /only:lex", "output": "lex: archetypes Carl Jung collective unconscious\nlex: archetypes significance literature psychology"}
+{"input": "machine learning recent news November /only:lex", "output": "lex: machine learning recent news November tutorial\nlex: machine learning recent news November guide\nlex: machine learning recent news November examples"}
+{"input": "Shopify recent news October /only:vec", "output": "vec: learn about Shopify recent news October\nvec: complete Shopify recent news October reference"}
+{"input": "AWS latest version release /only:vec", "output": "vec: complete AWS latest version release reference\nvec: understanding AWS latest version release"}
+{"input": "budget backpacking europe /only:lex", "output": "lex: budget backpacking europe documentation\nlex: budget backpacking europe guide\nlex: budget backpacking europe examples"}
+{"input": "what are the elements of classical music? /only:hyde", "output": "hyde: Classical music is built on melody (a sequence of notes forming a theme), harmony (chords supporting the melody), rhythm (the timing and pattern of notes), dynamics (volume changes), and form (the structure, such as sonata, rondo, or theme and variations)."}
+{"input": "electronics soldering guide /only:vec", "output": "vec: learn about electronics soldering guide\nvec: guide for electronics soldering guide"}
+{"input": "balance sheet basics /only:lex", "output": "lex: balance sheet basics guide\nlex: balance sheet basics tutorial\nlex: balance sheet basics examples"}
+{"input": "soccer formations tactics /only:vec", "output": "vec: complete soccer formations tactics reference\nvec: understanding soccer formations tactics"}
+{"input": "how to improve interpersonal skills /only:vec", "output": "vec: what are effective ways to improve interpersonal and communication skills?\nvec: how can someone develop better listening, empathy, and social skills in personal and professional settings?"}
+{"input": "enum class /only:lex", "output": "lex: enum class C++ Java strongly typed\nlex: enum class Python enumeration members\nlex: enum class scoped enumeration"}
+{"input": "how to prepare for a long hike /only:hyde", "output": "hyde: Train by walking with a loaded pack for progressively longer distances over 4-6 weeks. Pack the ten essentials: navigation, sun protection, insulation, illumination, first aid, fire, tools, nutrition, hydration, and shelter. Check the weather forecast and file a trip plan with someone you trust."}
+{"input": "latest findings in climate science /only:vec", "output": "vec: what are the most recent scientific findings about climate change in 2025-2026\nvec: what do the latest climate science studies reveal about global warming trends"}
+{"input": "python decorators explained /only:lex", "output": "lex: python decorator function\nlex: python @ decorator syntax\nlex: python wrapper decorator"}
+{"input": "awk command examples /only:vec", "output": "vec: how to use awk for text processing and extracting columns from files\nvec: what are common awk patterns and commands for parsing structured text"}
+{"input": "what are the building blocks of life /only:hyde", "output": "hyde: The building blocks of life are four types of organic molecules: proteins (made from amino acids), nucleic acids (DNA and RNA from nucleotides), carbohydrates (sugars and polysaccharides), and lipids (fats and phospholipids). These molecules self-assemble into cells, the basic unit of all living organisms."}
+{"input": "largest countries by area /only:vec", "output": "vec: understanding largest countries by area\nvec: complete largest countries by area reference"}
+{"input": "where to find budget travel tips /only:hyde", "output": "hyde: To travel on a budget, book flights midweek, use fare comparison tools like Google Flights or Skyscanner, stay in hostels or use house-sitting platforms, and eat at local markets instead of tourist restaurants."}
+{"input": "how to choose a daycare? /only:lex", "output": "lex: daycare choose selection criteria childcare\nlex: daycare center evaluation safety ratio"}
+{"input": "how to make homemade pizza /only:lex", "output": "lex: homemade pizza dough recipe\nlex: pizza from scratch oven toppings\nlex: make pizza dough sauce crust"}
+{"input": "mandarin tones guide /only:lex", "output": "lex: mandarin tones guide guide\nlex: mandarin tones guide best practices\nlex: mandarin tones guide examples"}
+{"input": "what is the concept of original sin /only:hyde", "output": "hyde: Original sin is the Christian doctrine that humanity inherited a sinful nature from Adam and Eve's disobedience in the Garden of Eden. Augustine of Hippo formalized the teaching, arguing that all humans are born in a state of sin, redeemable only through divine grace."}
+{"input": "art class /only:vec", "output": "vec: where can I find art classes for beginners to learn painting or drawing\nvec: what types of art classes are available online and in person for adults"}
+{"input": "what is the veil of ignorance /only:hyde", "output": "hyde: The veil of ignorance is a thought experiment by John Rawls in A Theory of Justice (1971). It asks people to choose principles of justice from an \"original position\" where they don't know their own race, gender, wealth, or abilities. Rawls argues this produces fair, impartial rules."}
+{"input": "thailand /only:hyde", "output": "hyde: Thailand is a Southeast Asian country known for tropical beaches, ornate temples, and rich cuisine. Bangkok is the capital. Popular destinations include Chiang Mai, Phuket, and the islands of Koh Samui and Phi Phi. Thai food staples include pad thai, green curry, and tom yum soup."}
+{"input": "how to make scientific presentations engaging /only:vec", "output": "vec: how can scientists make their research presentations more engaging and accessible\nvec: what techniques improve the delivery and visual design of scientific talks"}
+{"input": "spice combinations guide /only:lex", "output": "lex: spice combinations guide documentation\nlex: spice combinations guide tutorial\nlex: spice combinations guide examples"}
+{"input": "what is the function of dialogue? /only:vec", "output": "vec: what purpose does dialogue serve in communication and storytelling\nvec: how does dialogue function in literature and everyday interaction"}
+{"input": "how do philosophical arguments work /only:lex", "output": "lex: philosophical arguments logic premises conclusion\nlex: philosophical reasoning deductive inductive"}
+{"input": "what is the ethics of war /only:hyde", "output": "hyde: Just war theory establishes criteria for morally permissible warfare. Jus ad bellum (right to go to war) requires just cause, legitimate authority, right intention, last resort, proportionality, and reasonable chance of success. Jus in bello (right conduct in war) requires distinction between combatants and civilians and proportional use of force."}
+{"input": "what are creative portrait ideas? /only:lex", "output": "lex: creative portrait photography ideas techniques\nlex: portrait photo ideas poses lighting creative"}
+{"input": "how to lose weight fast? /only:hyde", "output": "hyde: Safe weight loss is 1-2 pounds per week through a calorie deficit of 500-1000 calories daily. Combine a protein-rich diet with strength training and cardio. Avoid crash diets\u2014they cause muscle loss and metabolic slowdown. Drink water, sleep 7-9 hours, and track food intake for accountability."}
+{"input": "how to choose the right camera /only:hyde", "output": "hyde: Decide what you'll shoot most: landscapes, portraits, video, or street photography. Mirrorless cameras are lighter with faster autofocus, while DSLRs offer longer battery life and more lens options. Key specs to compare: sensor size (full-frame vs APS-C), megapixels, autofocus points, and video capabilities. Budget $500-1000 for a capable starter body."}
+{"input": "how to encourage siblings to get along? /only:hyde", "output": "hyde: Give each child one-on-one time to reduce competition for attention. Avoid comparing siblings or labeling them (\"the smart one\"). Teach conflict resolution: help them express feelings with \"I\" statements and find compromises. Praise cooperation when you see it. Set clear family rules about physical aggression and name-calling."}
+{"input": "Renaissance Italy Florence /only:hyde", "output": "hyde: The Renaissance began in Florence around 1400 due to wealth from banking and trade, political stability, and classical heritage. The Medici family, especially Lorenzo the Magnificent, patronized artists like Leonardo, Michelangelo, and Botticelli. Florence's guilds, humanism from rediscovered Greek texts, and competition among city-states drove cultural innovation."}
+{"input": "how to evaluate scientific claims critically /only:lex", "output": "lex: evaluate scientific claims critical thinking\nlex: scientific literacy evidence evaluation peer review"}
+{"input": "tech fix /only:vec", "output": "vec: how to troubleshoot and fix common technology problems with computers and devices\nvec: what are basic tech fixes for common software and hardware issues"}
+{"input": "latest uses of bioinformatics in research /only:lex", "output": "lex: bioinformatics research applications 2025 2026\nlex: bioinformatics genomics proteomics computational biology"}
+{"input": "yoga poses beginners /only:vec", "output": "vec: learn about yoga poses beginners\nvec: guide for yoga poses beginners"}
+{"input": "where to find heirloom seed suppliers? /only:hyde", "output": "hyde: Top heirloom seed suppliers include Baker Creek Heirloom Seeds, Seed Savers Exchange, and Johnny's Selected Seeds. Baker Creek offers over 1,800 open-pollinated varieties with free shipping. Seed Savers Exchange is a nonprofit dedicated to preserving rare heirloom varieties through their seed bank and catalog."}
+{"input": "what is the composition of the earth's atmosphere /only:hyde", "output": "hyde: Earth's atmosphere is composed of 78.09% nitrogen (N\u2082), 20.95% oxygen (O\u2082), 0.93% argon (Ar), and 0.04% carbon dioxide (CO\u2082). Trace gases include neon, helium, methane, krypton, and water vapor (0-4% depending on humidity). The atmosphere extends roughly 480 km above the surface and is divided into five layers: troposphere, stratosphere, mesosphere, thermosphere, and exosphere."}
+{"input": "event sourcing pattern /only:vec", "output": "vec: what is event sourcing and how does it differ from traditional crud data storage\nvec: how do you implement event sourcing and what are its benefits and challenges"}
+{"input": "Spanish Conquest Americas /only:vec", "output": "vec: how did spanish conquistadors conquer the aztec and inca empires\nvec: what factors enabled spain to colonize the americas so rapidly in the 16th century"}
+{"input": "http client /only:vec", "output": "vec: how to make HTTP requests using an HTTP client library\nvec: which HTTP client libraries are available for making API calls in different languages"}
+{"input": "what are the sacred texts of judaism /only:vec", "output": "vec: what are the main sacred texts and scriptures in the Jewish religious tradition\nvec: what is the Torah and what other texts are considered holy in Judaism"}
+{"input": "endangered species list /only:vec", "output": "vec: learn about endangered species list\nvec: guide for endangered species list"}
+{"input": "how are glaciers formed /only:lex", "output": "lex: glacier formation process ice\nlex: glaciers formed snow compaction accumulation"}
+{"input": "what is content marketing /only:vec", "output": "vec: what is content marketing and how does it attract customers\nvec: how do businesses use content marketing to drive traffic and build trust"}
+{"input": "what is cycling commute? /only:hyde", "output": "hyde: Cycling commute refers to using a bicycle as your primary transportation to and from work. Bike commuters typically ride 3-15 miles each way, saving on fuel costs while getting daily exercise. Many cities now have protected bike lanes and bike-share programs."}
+{"input": "what is the importance of spiritual leadership? /only:lex", "output": "lex: spiritual leadership organizations values\nlex: spiritual leadership workplace meaning purpose"}
+{"input": "what is interfaith dialogue? /only:hyde", "output": "hyde: Interfaith dialogue is the cooperative interaction between people of different religious traditions, aimed at mutual understanding rather than conversion. Organizations like the Parliament of the World's Religions bring together leaders from Christianity, Islam, Judaism, Hinduism, Buddhism, and others to discuss shared values and address social issues."}
+{"input": "what is the role of enzymes in digestion /only:vec", "output": "vec: how do enzymes help break down food during the digestive process\nvec: what role do specific enzymes like amylase and protease play in digestion"}
+{"input": "Aztec Empire civilization /only:lex", "output": "lex: aztec empire civilization\nlex: aztec tenochtitlan mexico\nlex: aztec history mesoamerica"}
+{"input": "where to find landscaping stones? /only:hyde", "output": "hyde: Landscaping stones can be purchased from home improvement stores like Home Depot and Lowe's, local stone yards, and quarries. For bulk orders, landscape supply companies deliver directly. River rock, flagstone, and pea gravel are popular choices for garden paths and borders."}
+{"input": "how to stay informed about politics /only:hyde", "output": "hyde: Read multiple news sources across the political spectrum: AP News and Reuters for wire reporting, then compare coverage from different outlets. Subscribe to newsletters like The Morning (NYT) or Axios AM. Follow legislative trackers like Congress.gov. Attend local government meetings and candidate forums."}
+{"input": "latest GitHub updates /only:lex", "output": "lex: latest GitHub updates guide\nlex: latest GitHub updates documentation\nlex: latest GitHub updates best practices"}
+{"input": "how tourism affects local cultures /only:vec", "output": "vec: what are the positive and negative effects of tourism on local cultural traditions and communities\nvec: how does mass tourism change the customs, language, and daily life of host communities"}
+{"input": "how to find a reliable realtor /only:lex", "output": "lex: find reliable realtor real estate agent\nlex: choosing trustworthy real estate agent"}
+{"input": "monorepo vs polyrepo /only:vec", "output": "vec: what are the tradeoffs between using a monorepo versus multiple repositories\nvec: when does a monorepo make sense and what tools help manage large monorepos"}
+{"input": "habit formation science /only:lex", "output": "lex: habit formation science examples\nlex: habit formation science tutorial\nlex: habit formation science documentation"}
+{"input": "how to conserve energy in the office? /only:hyde", "output": "hyde: Switch to LED lighting and install occupancy sensors in conference rooms and restrooms. Set computers to sleep mode after 10 minutes of inactivity. Use smart power strips to eliminate phantom loads. Set thermostats to 68\u00b0F in winter and 76\u00b0F in summer. These measures typically reduce office energy use by 20-30%."}
+{"input": "what is a mathematical model /only:hyde", "output": "hyde: A mathematical model uses equations and variables to represent a real-world system. For example, the SIR model uses differential equations to predict infectious disease spread: dS/dt = -\u03b2SI, dI/dt = \u03b2SI - \u03b3I, dR/dt = \u03b3I. Models are validated by comparing predictions against observed data and refined iteratively."}
+{"input": "unit test vs integration test /only:lex", "output": "lex: unit test integration test difference\nlex: testing pyramid unit integration e2e\nlex: unit test isolation mocking"}
+{"input": "what changed in Python 2026 /only:vec", "output": "vec: guide for what changed in Python 2026\nvec: learn about what changed in Python 2026"}
+{"input": "how do philosophers explore the nature of reality /only:vec", "output": "vec: how have philosophers historically explored and debated the nature of reality and existence?\nvec: what are the main metaphysical positions on whether reality is fundamentally material, mental, or something else?"}
+{"input": "what are the teachings of the baha'i faith? /only:vec", "output": "vec: what are the core beliefs and teachings of the Baha'i faith\nvec: what did Baha'u'llah teach about unity, equality, and world peace"}
+{"input": "what is compositional balance? /only:hyde", "output": "hyde: Compositional balance refers to the distribution of visual weight within an image or artwork. Symmetrical balance places equal elements on both sides of a central axis, while asymmetrical balance uses contrasting elements \u2014 such as a large shape offset by a smaller, brighter one \u2014 to create dynamic equilibrium."}
+{"input": "how does the body maintain homeostasis /only:vec", "output": "vec: what mechanisms does the human body use to maintain internal stability\nvec: how do feedback loops help regulate body temperature and blood sugar levels"}
+{"input": "database transaction isolation levels /only:hyde", "output": "hyde: Isolation levels from weakest to strongest: Read Uncommitted (dirty reads possible), Read Committed (sees only committed data, default in PostgreSQL), Repeatable Read (no non-repeatable reads), Serializable (no phantom reads, full isolation). Higher isolation = more locking = lower concurrency. Choose based on consistency needs vs performance."}
+{"input": "what to pack in a hospital bag for labor? /only:hyde", "output": "hyde: Hospital bag essentials for labor: ID and insurance card, birth plan, comfortable robe or gown, slippers, toiletries, phone charger, going-home outfit for you and baby, car seat, nursing bra, newborn diapers, snacks, and a pillow from home."}
+{"input": "how to learn about native american culture /only:hyde", "output": "hyde: Visit the National Museum of the American Indian (Smithsonian) or local tribal cultural centers. Read works by Native authors like Joy Harjo, Tommy Orange, and Robin Wall Kimmerer. Attend powwows and cultural events when open to the public. Learn which tribal nations are indigenous to your area."}
+{"input": "what is burnout? /only:lex", "output": "lex: burnout syndrome workplace exhaustion\nlex: burnout symptoms causes recovery"}
+{"input": "dependency injection benefits /only:hyde", "output": "hyde: Dependency injection provides dependencies from outside rather than creating them internally. Class receives DatabaseService via constructor instead of instantiating it. Benefits: loose coupling, easy testing with mocks, flexible configuration. Instead of new EmailService(), inject interface IEmailService\u2014swap implementations without changing consumer code."}
+{"input": "what is the role of media in politics /only:vec", "output": "vec: what role does the media play in shaping political discourse and public opinion\nvec: how does news coverage and media bias influence political outcomes and democracy"}
+{"input": "what are plasmids /only:lex", "output": "lex: plasmid DNA circular extrachromosomal\nlex: plasmid bacteria gene transfer cloning\nlex: plasmid vector molecular biology"}
+{"input": "how to engage in civil political discussions /only:hyde", "output": "hyde: Start by listening to understand, not to rebut. Ask questions like \"What experiences led you to that view?\" Avoid personal attacks and generalizations. Find common ground before addressing differences. Use \"I\" statements instead of \"you always\" accusations. Accept that changing minds takes time and repeated respectful engagement."}
+{"input": "regex lookahead lookbehind /only:hyde", "output": "hyde: Lookahead (?=pattern) matches a position followed by pattern without consuming it. Negative lookahead (?!pattern) matches where pattern doesn't follow. Lookbehind (?<=pattern) matches a position preceded by pattern. Example: \\d+(?= dollars) matches numbers followed by 'dollars'."}
+{"input": "how to diversify investment portfolio /only:hyde", "output": "hyde: Diversify across asset classes: stocks, bonds, real estate, and commodities. Within stocks, spread across sectors (tech, healthcare, energy) and geographies (US, international, emerging markets). Use index funds or ETFs for broad exposure. A common allocation is 60% stocks, 30% bonds, 10% alternatives, adjusted by age and risk tolerance."}
+{"input": "recent GitHub changes 2026 /only:lex", "output": "lex: recent GitHub changes 2026 tutorial\nlex: recent GitHub changes 2026 examples\nlex: recent GitHub changes 2026 guide"}
+{"input": "Sentry error tracking /only:vec", "output": "vec: understanding Sentry error tracking\nvec: learn about Sentry error tracking"}
+{"input": "what is the principle of utility? /only:vec", "output": "vec: what is the principle of utility in utilitarian ethics as defined by Bentham and Mill\nvec: how does the utilitarian principle of utility evaluate actions based on their consequences for overall happiness"}
+{"input": "portrait photography tips /only:vec", "output": "vec: what are the best tips for taking professional-quality portrait photographs?\nvec: how should you set up lighting, posing, and camera settings for portrait photography?"}
+{"input": "what are the core practices of the bah\u00e1'\u00ed faith? /only:lex", "output": "lex: Bah\u00e1'\u00ed faith core practices worship\nlex: Bah\u00e1'\u00ed religion prayer fasting principles"}
+{"input": "how to find emotional support /only:vec", "output": "vec: where can someone find emotional support during difficult times or mental health challenges\nvec: what resources are available for people seeking emotional support and counseling"}
+{"input": "soccer training drills /only:vec", "output": "vec: what are effective soccer training drills for improving skills and fitness\nvec: which soccer drills help players improve dribbling, passing, and shooting"}
+{"input": "how to participate in a protest /only:hyde", "output": "hyde: Know your rights: the First Amendment protects peaceful assembly on public property. Bring water, snacks, a phone charger, and ID. Write an emergency contact number on your arm. Stay with a buddy and agree on a meeting point. Wear comfortable shoes and weather-appropriate clothing. If tear gas is used, move upwind. Document police interactions by filming at a safe distance."}
+{"input": "what is the importance of peer review /only:lex", "output": "lex: peer review importance scientific publishing\nlex: peer review process academic research"}
+{"input": "idempotency api design /only:hyde", "output": "hyde: Idempotent operations produce the same result regardless of how many times called. GET, PUT, DELETE are naturally idempotent. POST needs idempotency keys: client sends unique key, server stores result, returns cached result on retry. Store keys with TTL (24h). Critical for payment APIs\u2014prevents double charges on network retry."}
+{"input": "AWS Lambda functions setup /only:lex", "output": "lex: AWS Lambda functions setup documentation\nlex: AWS Lambda functions setup examples\nlex: AWS Lambda functions setup tutorial"}
+{"input": "how to negotiate a business deal /only:hyde", "output": "hyde: Prepare by researching the other party's priorities and constraints. Define your BATNA (best alternative to a negotiated agreement) and walk-away point. Open with an ambitious but defensible anchor. Listen more than you talk. Focus on interests, not positions, to find creative win-win solutions."}
+{"input": "how do ethical theories apply to social issues /only:vec", "output": "vec: how are ethical theories like utilitarianism and deontology applied to real-world social issues?\nvec: what ethical frameworks do philosophers use to analyze problems like poverty, inequality, and healthcare?"}
+{"input": "S3 bucket policy /only:hyde", "output": "hyde: S3 bucket policies are resource-based JSON policies attached to buckets. Grant public read: {\"Statement\":[{\"Effect\":\"Allow\",\"Principal\":\"*\",\"Action\":\"s3:GetObject\",\"Resource\":\"arn:aws:s3:::bucket/*\"}]}. IAM policies attach to users/roles. Use bucket policies for cross-account access, IAM for user-specific permissions. Block public access settings override policies."}
+{"input": "what is virtue signaling? /only:vec", "output": "vec: what does virtue signaling mean and how is the term used in political and social discourse?\nvec: how do people use virtue signaling to publicly express moral values without substantive action?"}
+{"input": "what is the significance of the sacred tree in various faiths? /only:hyde", "output": "hyde: Sacred trees appear across religions: the Bodhi tree where Buddha attained enlightenment, the Tree of Life in Genesis, Yggdrasil in Norse mythology connecting the nine worlds, and the banyan in Hinduism symbolizing eternal life. Trees represent growth, connection between earth and heaven, and renewal."}
+{"input": "list sort /only:lex", "output": "lex: sort list programming algorithm\nlex: list sort Python Java ascending descending\nlex: array sorting methods comparison"}
+{"input": "what makes a good thriller novel? /only:lex", "output": "lex: thriller novel elements writing techniques\nlex: good thriller pacing suspense plot twists"}
+{"input": "what is the significance of community in spirituality? /only:hyde", "output": "hyde: Spiritual communities provide shared worship, accountability, and mutual support that deepen individual faith. In Christianity, the church body gathers for fellowship; in Buddhism, the sangha is one of the Three Jewels; in Judaism, a minyan of ten is required for communal prayer. Communal practice reinforces commitment and provides belonging."}
+{"input": "latest Python updates /only:vec", "output": "vec: how to latest Python updates\nvec: complete latest Python updates reference"}
+{"input": "how to pursue a career in scientific research /only:lex", "output": "lex: scientific research career path academia\nlex: career scientist PhD research position"}
+{"input": "where to buy organic seeds? /only:lex", "output": "lex: buy organic seeds online garden\nlex: organic seed suppliers heirloom non-GMO"}
+{"input": "industrial revolution inventions /only:vec", "output": "vec: how to industrial revolution inventions\nvec: guide for industrial revolution inventions"}
+{"input": "what is an elevator pitch /only:hyde", "output": "hyde: An elevator pitch is a concise, 30-60 second summary of who you are and what you offer. Structure it as: hook (attention-grabbing opening), problem you solve, your solution, and a call to action. Practice until it sounds conversational, not rehearsed."}
+{"input": "how to meditate for beginners /only:hyde", "output": "hyde: Sit comfortably with your back straight. Close your eyes and focus on your breath\u2014notice each inhale and exhale. When thoughts arise, gently return attention to your breathing without judgment. Start with 5 minutes daily and gradually increase. Consistency matters more than duration."}
+{"input": "where to find datasets for scientific research /only:hyde", "output": "hyde: Public research datasets are available from repositories such as Kaggle, the UCI Machine Learning Repository, NASA's Open Data Portal, NOAA Climate Data, and institutional data archives like Harvard Dataverse and Zenodo."}

finetune/prepare_v4_dataset.py

@@ -0,0 +1,107 @@
+# /// script
+# requires-python = ">=3.10"
+# dependencies = []
+# ///
+"""Prepare v4 dataset: high-quality expansions + /only: variants."""
+
+import json
+import random
+from pathlib import Path
+
+def to_chat_format(query: str, output: str) -> dict:
+    """Convert input/output to chat format with /no_think."""
+    # For /only: queries, keep the suffix in the prompt
+    prompt = f"/no_think Expand this search query: {query}"
+    
+    text = f"<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n{output}<|im_end|>\n"
+    
+    messages = [
+        {"role": "user", "content": prompt},
+        {"role": "assistant", "content": output}
+    ]
+    
+    return {"text": text, "messages": messages}
+
+
+def load_jsonl(path: Path) -> list[dict]:
+    """Load JSONL file."""
+    data = []
+    with open(path) as f:
+        for line in f:
+            line = line.strip()
+            if line:
+                data.append(json.loads(line))
+    return data
+
+
+def main():
+    data_dir = Path("data")
+    
+    # High-quality sources
+    sources = [
+        ("qmd_expansion_v2.jsonl", "v2"),
+        ("qmd_expansion_handcrafted.jsonl", "handcrafted"),
+        ("qmd_only_variants.jsonl", "only"),
+    ]
+    
+    all_examples = []
+    stats = {}
+    
+    for filename, label in sources:
+        path = data_dir / filename
+        if not path.exists():
+            print(f"  Skipping {filename} (not found)")
+            continue
+        
+        raw = load_jsonl(path)
+        converted = []
+        
+        for item in raw:
+            query = item.get("input", "")
+            output = item.get("output", "")
+            if query and output:
+                converted.append(to_chat_format(query, output))
+        
+        all_examples.extend(converted)
+        stats[label] = len(converted)
+        print(f"  {label}: {len(converted)} examples")
+    
+    # Shuffle
+    random.seed(42)
+    random.shuffle(all_examples)
+    
+    # Split 90/10
+    split_idx = int(len(all_examples) * 0.9)
+    train = all_examples[:split_idx]
+    val = all_examples[split_idx:]
+    
+    # Write output
+    out_dir = data_dir / "train_v4"
+    out_dir.mkdir(exist_ok=True)
+    
+    with open(out_dir / "train.jsonl", "w") as f:
+        for ex in train:
+            f.write(json.dumps(ex) + "\n")
+    
+    with open(out_dir / "val.jsonl", "w") as f:
+        for ex in val:
+            f.write(json.dumps(ex) + "\n")
+    
+    # Dataset info
+    info = {
+        "dataset_name": "qmd-query-expansion-v4",
+        "train_samples": len(train),
+        "val_samples": len(val),
+        "sources": stats,
+    }
+    with open(out_dir / "dataset_info.json", "w") as f:
+        json.dump(info, f, indent=2)
+    
+    print(f"\n✓ Dataset prepared in {out_dir}/")
+    print(f"  Train: {len(train)}")
+    print(f"  Val: {len(val)}")
+    print(f"  Total: {len(all_examples)}")
+
+
+if __name__ == "__main__":
+    main()