flint-flat-adaptive-qwen3.5-4b

# flint-flat-adaptive-qwen3.5-4b Compressed ("caveman") reasoning traces for SFT — the **flat-adaptive** variant of the flint reasoning-compression pipeline. Converted from verified self-distilled traces by **Qwen/Qwen3.5-4B**, policy `policy/1.1`, template `caveman_convert/2.0`. Flat compression, level adaptive to trace length (short = heavy, long = light), wind-down tail preserved verbatim (policy 1.1). Each row: `input`, `reasoning` (compressed trace), `answer` (carried verbatim from the source), `domain`, `verified`, `difficulty`, and `meta` with per-row provenance (converter/segmenter model, span labels + actions for section-aware rows, token counts raw vs compressed). - Rows: **332** — domains: {"general": 110, "math": 106, "creative": 105, "code": 11} - Mean compression ratio (caveman/raw tokens): **0.35** - Content-preservation linters (numbers, negations, byte-identical code blocks, repetition) reject unsafe conversions; rejects are excluded. Part of the caveman reasoning-compression ablation study. ## Build manifest ```json { "created_at": "2026-07-09T12:09:04+00:00", "git_sha": "cb1eaae9ddc9ba9acd8c5554488d2fdf1d3cf52e", "variant": "flat-adaptive", "config": { "mode": "flat", "level": "adaptive" }, "counts": { "ok": 332, "rejected": 316, "failed": 0 }, "convert_model": "Qwen/Qwen3.5-4B", "segment_model": null, "template_version": "caveman_convert/2.0", "policy_version": "policy/1.1", "segment_template_version": null, "tokenizer": "Qwen/Qwen3.5-4B", "input_rows": 648 } ```