marcodsn/crucible-sft-gemma-4-12b-it-mini

crucible-sft-gemma-4-12b-it-mini

Self-distilled SFT dataset of verified reasoning traces from unsloth/gemma-4-12b-it, built by the reasoning-compression crucible pipeline: k-sample generation on a decontaminated prompt pool, inline verification (symbolic math / sandboxed code tests), difficulty banding via solve rate, and loop-detector filtering on the chosen trace.

Each row: prompt, reasoning (a verified-correct thinking trace when the domain is verifiable), response, domain, verified, difficulty ({solve_rate, k}), sampling params in gen.

• —Rows: 824 — domains: {"math": 196, "creative": 308, "general": 259, "code": 61}
• —Prompts are content-addressed (prompt_id = sha256 of prompt text) and decontaminated against GSM8K, MATH-500, HumanEval, MBPP, MMLU-Pro.
• —A stratified held-out split was carved before generation and is not included here.

Part of the caveman reasoning-compression ablation study (see the flint datasets built from this file and the trained comparison arms).

Build manifest

{
  "created_at": "2026-07-10T23:23:51+00:00",
  "git_sha": "9b2ffbbeeddc3f9094805e1ea9a07d2dedaf2eb7",
  "recipe": "flint-ablation-v1-mini",
  "recipe_def": {
    "target_size": 1100,
    "shares": {
      "math": 0.3,
      "code": 0.1,
      "general": 0.32,
      "creative": 0.28
    },
    "candidate_caps": {
      "math": 700,
      "code": 250,
      "general": 400,
      "creative": 350
    },
    "solve_band": [
      0.2,
      0.8
    ],
    "k": 4,
    "oversample": 2.2,
    "lang": "en"
  },
  "raw_rows": 1620,
  "selected_rows": 824,
  "domains": {
    "math": 196,
    "creative": 308,
    "general": 259,
    "code": 61
  },
  "verified": {
    "True": 257,
    "None": 567
  },
  "degenerate_in_raw": 86,
  "gen_model": "unsloth/gemma-4-12b-it"
}