Datasetpublic

hellouniverse/MiMo-SFT2-math500-responses

MiMo SFT2 — MATH-500 responses (128× sampling, temp 0.6) Model responses generated with an SFT2 MiMo-7B model (tequila3009/sft2_mimo, weights under the sft2_mimo/ subdir) on the MATH-500 problem set. Generation setup Model SFT2 MiMo-7B — tequila3009/sft2_mimo (sft2_mimo/) Dataset MATH-500 — 500 problems Samples per problem 128 Total responses 64,000 Temperature 0.6 top_p 0.95 top_k -1 (disabled) max_tokens 16384 Engine vLLM, TP=8 on 8×… See the full description on the dataset page: https://huggingface.co/datasets/hellouniverse/MiMo-SFT2-math500-responses.

sourceHugging Faceapache-2.0updated 3mo agoView on Hugging Face

0likes12downloads

Dataset Card

MiMo SFT2 — MATH-500 responses (128× sampling, temp 0.6)

Model responses generated with an SFT2 MiMo-7B model (`tequila3009/sft2_mimo`, weights under the sft2_mimo/ subdir) on the MATH-500 problem set.

Generation setup


Model	SFT2 MiMo-7B — `tequila3009/sft2_mimo` (`sft2_mimo/`)
Dataset	MATH-500 — 500 problems
Samples per problem	128
Total responses	64,000
Temperature	0.6
top_p	0.95
top_k	-1 (disabled)
max_tokens	16384
Engine	vLLM, TP=8 on 8× A100

Each problem is wrapped in a natural-language chain-of-thought prompt that asks the model to reason step by step and put the final answer in \boxed{...}.

File

—sft2_mimo_math500_responses_t0.6_20260707_134307.json (~657 MB)

Structure

json

{
  "model": "sft2_mimo",
  "dataset": "math500.jsonl",
  "sampling_params": {"temperature": 0.6, "top_p": 0.95, "top_k": -1, "max_tokens": 16384},
  "total_questions": 500,
  "responses_per_question": 128,
  "generation_time": 34920.0,
  "results": [
    {"question": "<problem text>", "responses": ["<sample 1>", "... 128 samples ..."]}
  ]
}

results is aligned to the MATH-500 order; results[i]["responses"] holds the 128 samples for problem i.

Notes

—All 64,000 responses are non-empty (0 empty completions).
—Responses are raw model text (reasoning + \boxed{} answer), not verified for correctness.

Loading

python

import json
data = json.load(open("sft2_mimo_math500_responses_t0.6_20260707_134307.json"))
print(data["total_questions"], data["responses_per_question"])   # 500 128
q0 = data["results"][0]
print(q0["question"])
print(q0["responses"][0])