Dataset Viewer
Auto-converted to Parquet Duplicate
messages
listlengths
2
4
outputs
dict
[ { "content": "You are a concise assistant.", "role": "system" }, { "content": "Give me one sentence about the moon.", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user asks: \"Give me one sentence about the moon.\" We need to produce a single sentence about the moon. Concise. Should be one sentence. Provide.assistantfinalThe moon, Earth's only natural satellite, reflects sunlight and influences our tides, calendars, and cultu...
[ { "content": "You explain math clearly.", "role": "system" }, { "content": "What is 7 * 8?", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user asks a simple multiplication. Provide answer: 56. Also maybe explain.assistantfinal7 × 8 = 56. \n\nMultiplying 7 by 8 means adding 7 together eight times:\n\n7 + 7 + 7 + 7 + 7 + 7 + 7 + 7 = 56. \n\nSo the product of 7 and 8 is **56**." }
[ { "content": "You are a coding tutor.", "role": "system" }, { "content": "How do I reverse a list in Python?", "role": "user" }, { "content": "Use list[::-1] or list.reverse().", "role": "assistant" }, { "content": "Show a short example with list[::-1].", "role": "user" ...
{ "reasoning_content": null, "text": "analysisWe need to comply with policy. The user wants a short example with list[::-1] to reverse a list. That's fine. Provide code snippet. No disallowed content. Provide explanation.assistantfinalSure! The slice syntax `list[::-1]` creates a new list that contains the elements...
[ { "content": "You are a travel planner.", "role": "system" }, { "content": "Give me two things to do in Seattle.", "role": "user" }, { "content": "Visit Pike Place and walk Discovery Park.", "role": "assistant" }, { "content": "Now make it indoor activities only.", "role"...
{ "reasoning_content": null, "text": "analysisUser wants two indoor activities in Seattle. Provide two suggestions. Probably include details.assistantfinalHere are two great **indoor** activities you can enjoy in Seattle:\n\n| # | Activity | Why It’s Worth It | How to Get There | Tips & Extras |\n|---|----------|--...
[ { "content": "You are a product naming assistant.", "role": "system" }, { "content": "Suggest a name for a habit tracker app.", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisWe need to suggest a name for a habit tracker app. The user just asks \"Suggest a name for a habit tracker app.\" We should respond with suggestions. Possibly multiple options, maybe a short list, with brief explanation. Should be creative. No disallowed content. Provid...
[ { "content": "You summarize text.", "role": "system" }, { "content": "Summarize: The cat sat on the mat all day.", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user asks to summarize: \"The cat sat on the mat all day.\" It's already a short sentence. Summarize: \"A cat stayed on a mat all day.\" Or \"The cat spent the whole day on the mat.\" Provide a concise summary.assistantfinalThe cat spent the entire day on the mat." ...
[ { "content": "You are a cooking helper.", "role": "system" }, { "content": "How long should I boil an egg for soft yolk?", "role": "user" }, { "content": "About 6 to 7 minutes in gently boiling water.", "role": "assistant" }, { "content": "What if I want medium yolk instead?"...
{ "reasoning_content": null, "text": "analysisUser asks about boiling eggs for different yolk consistencies. Provide cooking times, tips, method. Should be helpful. Provide steps. Also note altitude, egg size, starting temperature. Provide safety.assistantfinal**Boiling an Egg – How Long for a Medium‑Yolk Center?**...
[ { "content": "You rewrite text in plain language.", "role": "system" }, { "content": "Rewrite: Utilize this mechanism to facilitate outcomes.", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user wants a rewrite of \"Utilize this mechanism to facilitate outcomes.\" in plain language. So we need to rewrite in plain language. Likely: \"Use this method to help achieve results.\" Or \"Use this tool to get results.\" Provide a plain language version.assistan...
[ { "content": "You are a friendly Q&A bot.", "role": "system" }, { "content": "What is the capital of Japan?", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user asks: \"What is the capital of Japan?\" Straightforward: answer: Tokyo. Provide friendly tone.assistantfinalThe capital of Japan is **Tokyo**. 🌸" }
[ { "content": "You answer with short bullet points.", "role": "system" }, { "content": "List two benefits of daily walking.", "role": "user" } ]
{ "reasoning_content": null, "text": "analysisThe user asks: \"List two benefits of daily walking.\" The developer says: \"You answer with short bullet points.\" So we need to respond with short bullet points, two items. Should be concise. Provide two benefits. Use bullet points.assistantfinal- Improves cardiovascu...

Generated Responses Dataset

This dataset contains generated responses for prompts from tytodd/test-job-dataset-input.

Generation Details

Sampling Parameters

  • Temperature: 0.7
  • Top P: 0.8
  • Top K: 20
  • Min P: 0.0
  • Max Tokens: 16384
  • Repetition Penalty: 1.0

Hardware Configuration

  • Tensor Parallel Size: 4
  • GPU Configuration: 4 GPU(s)

Dataset Structure

The dataset contains all columns from the source dataset plus:

  • outputs: The generated outputs dict from the model

Generation Script

Generated using the vLLM inference script from modaic/batch-vllm.

To reproduce this generation:

uv run https://huggingface.co/datasets/modaic/batch-vllm/resolve/main/generate_responses.py \
    tytodd/test-job-dataset-input \
    <output-dataset> \
    --model-id openai/gpt-oss-120b \
    --messages-column messages \
    --temperature 0.7 \
    --top-p 0.8 \
    --top-k 20 \
    --max-tokens 16384
Downloads last month
12