Datasets:
messages listlengths 2 2 | category stringclasses 29
values | subcategory stringlengths 2 41 |
|---|---|---|
[
{
"role": "user",
"content": "A client sent these instructions for a product blurb about their oat milk: \"Do not fail to avoid mentioning any competitor. It would not be unwelcome if you refrained from health claims. Never omit the shelf life, which is 9 days once opened. We would not want you to not inclu... | instruction_following | negation_spec |
[
{
"role": "user",
"content": "Evaluate a museum proposal to digitize and repatriate a colonial archive while retaining local copies. Address ownership, description, physical return, digital sovereignty, contested provenance, community-defined access, conservation resources, and the danger of treating digiti... | knowledge_qa | archival_reparations |
[
{
"role": "user",
"content": "Design simulation and rendering LOD for a city crowd. Preserve aggregate flow, identity of important agents, animation, collisions, schedules, audio, and network relevance while switching representations without visible pops."
},
{
"role": "assistant",
"content": "*... | game_development | crowd_lod |
[
{
"role": "user",
"content": "A live game changes a branching quest after thousands of players have partial progress. Design versioned quest state and migration rules for removed nodes, split objectives, claimed rewards, co-op divergence, rollback, and customer-support inspection."
},
{
"role": "ass... | game_development | quest_graph_migration |
[
{
"role": "user",
"content": "Calibrate a three-axis magnetometer with hard-iron bias, soft-iron distortion, axis misalignment, and temperature drift. Design orientation data collection, ellipsoid fitting, validation, and fusion checks without assuming the local magnetic field is constant indoors."
},
{... | knowledge_qa | magnetometer_calibration |
[
{
"role": "user",
"content": "Select three distinct rewards uniformly from ten candidates. Use a partial shuffle or sampled indices, inject the random source, reject requests larger than the pool, and test no duplicates."
},
{
"role": "assistant",
"content": "```python\nimport random\nfrom typin... | game_development | random_without_replacement |
[
{
"role": "user",
"content": "Extract the email thread below into one JSON object: {\"thread_subject\": string | null, \"participants\": [{\"name\": string, \"email\": string | null, \"side\": \"client\" | \"venue\" | \"unknown\"}], \"event\": {\"type\": string | null, \"date\": \"YYYY-MM-DD\" | null, \"gue... | instruction_following | json_extraction |
[
{
"role": "user",
"content": "A browser application needs SharedArrayBuffer but embeds third-party media and analytics. Design COOP and COEP deployment, inventory corp and CORS requirements for every subresource, isolate incompatible frames, and test popup and authentication flows affected by browsing-conte... | web_development | cross_origin_isolation |
[
{
"role": "user",
"content": "Reconstruct pressure–temperature–time history from mineral assemblages, zoning, reactions, inclusions, and geochronology. Account for disequilibrium, diffusion, polymetamorphism, fluid, and uncertainty in thermodynamic data."
},
{
"role": "assistant",
"content": "A ... | earth_science | metamorphic_path |
[
{
"role": "user",
"content": "Plan sample size from target effect, variance, error rates, allocation, attrition, and design effect. Explain why post-hoc power adds little beyond the confidence interval and why powering only for a large effect risks inconclusive evidence."
},
{
"role": "assistant",
... | research_methods | power_and_precision |
[
{
"role": "user",
"content": "Fit or compare exponential and logistic growth to short population data. Estimate parameters, inspect residuals, identify non-identifiability before saturation, and explain why a fitted carrying capacity may not be permanent."
},
{
"role": "assistant",
"content": "*... | math | population_growth_models |
[
{
"role": "user",
"content": "Solve 2 times 3^x=54 using logarithms or recognition, then solve 2 times 3^x=50 to three decimal places. Check both by substitution and state why logarithm arguments must be positive."
},
{
"role": "assistant",
"content": "**Answers: x = 3 exactly, and x ≈ 2.930 to ... | math | logarithms |
[
{
"role": "user",
"content": "Our @MainActor view model freezes the UI for 1-2 s when opening a conversation, and the hang detector confirms main-thread work. The code:\n\n@MainActor\nfinal class ConversationViewModel: ObservableObject {\n @Published var messages: [MessageRow] = []\n\n func load(conve... | mobile_development | swift_concurrency |
[
{
"role": "user",
"content": "Estimate the sample size needed to measure a population proportion with 95 percent margin of error at most 3 percentage points using the conservative variance choice. Derive the formula, round correctly, and discuss finite-population correction when the entire population has on... | math | sample_size |
[
{
"role": "user",
"content": "Use the ideal-gas law and Dalton's law to analyze a mixture collected over water. Correct for water vapor pressure, calculate dry-gas moles and partial pressures, and state assumptions about equilibrium, temperature, and nonideality."
},
{
"role": "assistant",
"cont... | chemistry | gas_mixture |
[
{
"role": "user",
"content": "Compare auxin, cytokinin, gibberellin, abscisic acid, and ethylene in plant growth and stress. Analyze interaction, transport, concentration, tissue, and timing rather than assigning each hormone one fixed effect."
},
{
"role": "assistant",
"content": "## Core compa... | biology | plant_hormone_network |
[
{
"role": "user",
"content": "Implement a transaction helper that commits only after callback success, rolls back on error or cancellation, preserves the primary exception if rollback fails, and prevents nested use from silently changing semantics."
},
{
"role": "assistant",
"content": "```pytho... | coding | transaction_helper |
[
{
"role": "user",
"content": "Develop a formal analysis of an unfamiliar portrait before researching its artist. Address composition, scale, gaze, gesture, light, color, material, setting, and viewer position; separate observations from iconographic hypotheses, then list contextual evidence needed to test t... | knowledge_qa | visual_culture |
[
{
"role": "user",
"content": "Explain what a database index is to someone who has only ever used spreadsheets, and describe one situation where adding an index actually makes things slower."
},
{
"role": "assistant",
"content": "A database index is like the tab dividers in a giant binder of spre... | explanation | explanation |
[
{
"role": "user",
"content": "Classify each error message below by layer: FRONTEND, BACKEND, DATABASE, or NETWORK. Output a JSON array of objects {\"n\": number, \"layer\": string, \"confidence\": \"high\" | \"medium\" | \"low\", \"evidence\": verbatim substring of at most 40 characters}. When symptoms span... | instruction_following | classification |
[
{
"role": "user",
"content": "Write GoogleTest tests for this LRU cache interface. Cover: eviction order after a `get` refreshes recency, `put` of an existing key updates value without evicting, capacity 1, capacity 0 behavior (decide and document what you expect), and that `get` on a missing key returns `s... | coding | cpp_testing |
[
{
"role": "user",
"content": "Implement a WidgetKit widget showing the next 3 tram departures for a saved stop. Requirements: a TimelineProvider whose timeline contains one entry per minute for the next 30 minutes so the countdown stays fresh without reloads, .after reload policy at the 30-minute mark, a pl... | mobile_development | ios_widgets |
[
{
"role": "user",
"content": "Epoch-by-epoch trace for a text classifier: epoch 1 train loss 0.61 val loss 0.58; epoch 3 train 0.42 val 0.45; epoch 5 train 0.30 val 0.39; epoch 7 train 0.22 val 0.37; epoch 9 train 0.15 val 0.38; epoch 11 train 0.11 val 0.43; epoch 13 train 0.08 val 0.52; epoch 15 train 0.06... | machine_learning | bias_variance |
[
{
"role": "user",
"content": "A two-stage machining cell: Stage 1 processes 50 units/hour, Stage 2 processes 40 units/hour. Inspection after Stage 2 fails 15% of units, and failed units must repeat Stage 2 once (rework always passes). Compute the effective capacity of the cell in good units per hour, identi... | economics_finance | operations |
[
{
"role": "user",
"content": "Two sides of a triangle are 7 and 10 with included angle 60 degrees. Use the law of cosines to find the third side, then compute area two ways and check whether the dimensions are consistent."
},
{
"role": "assistant",
"content": "**Answer:** The third side is c = √... | math | triangle_law_cosines |
[
{
"role": "user",
"content": "Write release notes for a process or product update covering new behavior, fixes, known limitations, compatibility, migration, rollback, and support. Separate user-visible effects from internal changes and avoid marketing every fix as an enhancement."
},
{
"role": "assi... | writing | change_log |
[
{
"role": "user",
"content": "Design the integrity and non-repudiation controls for a double-entry payment ledger deployed active-active in two regions with asynchronous replication. Failure modes to design against: silent modification of posted entries, replay of transfer commands during failover, divergen... | security_engineering | threat_modeling_design |
End of preview. Expand in Data Studio
Ox Alpha - 10k
10,005 single-turn prompts for text-response teacher generation
Each row carries id, category, subcategory
All data was gathered using stealth/ox-alpha via OpenRouter (reasoning effort high)
Topic distribution
| Category | Prompts | Share |
|---|---|---|
| Coding (incl. Go/Rust, C++/Java/C#, shell/CLI) | 944 | 9.4% |
| Knowledge QA | 900 | 9.0% |
| Logical reasoning & decisions | 734 | 7.3% |
| Web development | 720 | 7.2% |
| Game development | 720 | 7.2% |
| Three.js / browser 3D | 620 | 6.2% |
| Writing & communication | 585 | 5.8% |
| Math & quantitative reasoning | 575 | 5.7% |
| Mobile development | 340 | 3.4% |
| Instruction following & structured output | 311 | 3.1% |
| Creativity | 300 | 3.0% |
| Biology | 275 | 2.7% |
| Physics & astronomy | 250 | 2.5% |
| Security engineering (defensive) | 240 | 2.4% |
| Machine learning | 240 | 2.4% |
| Economics & finance | 240 | 2.4% |
| Law & policy | 240 | 2.4% |
| Health & medicine | 240 | 2.4% |
| Multilingual & translation | 240 | 2.4% |
| Education & tutoring | 235 | 2.3% |
| Chemistry | 225 | 2.2% |
| Humanities & social science | 200 | 2.0% |
| Data tasks (SQL, spreadsheets, wrangling) | 185 | 1.8% |
| Earth & environmental science | 150 | 1.5% |
| Engineering | 150 | 1.5% |
| Research methods | 75 | 0.7% |
| DevOps / SRE | 40 | 0.4% |
| Safety | 16 | 0.2% |
| Explanation | 15 | 0.1% |
| Total | 10,005 | 100% |
Difficulty
| Tier | Prompts | Share |
|---|---|---|
| Hard | 4,402 | 44.0% |
| Medium | 3,456 | 34.5% |
| Expert | 984 | 9.8% |
| Foundational | 965 | 9.6% |
| Mixed (curated seeds) | 198 | 2.0% |
Family
| Family | Prompts | Share |
|---|---|---|
| Ability (do a task) | 7,103 | 71.0% |
| Knowledge (explain/analyze) | 2,704 | 27.0% |
| Curated seed (original hand-picked) | 198 | 2.0% |
- Downloads last month
- -