Model card update (BLiMP recomputed after harness fix; data-attribution experiments) + READMEs for 4 experiment folders

card: honest wiki correction (byte_ppl 2.372 / BPB 1.246 / eff ~75.9); maintainer re-benchmark PR#78; ARC-Easy 47.94 exact-match

card: dedup 64M load-recipe (single clean block)

card: add 64M Qwen3 load-recipe (maintainer audit load-safety)

card: 64M #6 flagship, board-protocol verified (glint bare-prompt: eff 77.51 / ARC 47.94)

revert card to pre-64M: hold 64M #5 pending board-protocol (glint bare-prompt) re-measure

card: add 64M #5 flagship (eff 77.70, board-native recompute-verified)

card: document ARC-MIX 9.42B corpus (32M/64M runs) + v2 FineWeb-Edu-dominant blend plan

card: confirmed board entry (#16 32M / #21 16M, maintainer-verified); all models + results; ARC lever = value residuals (competitive intel); 64M A/B roadmap

card: Positioning CONFIRMED on board #16/#21 + Roadmap VR-ARC-lever/data-stack/64M-A/B

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Usage: reference in-repo train_gpt_ref.py and glint_parity_eval.py

Polished professional card: model-details, usage, evaluation-protocol, key-findings, roadmap, limitations

Add Usage section (merge best-of-both cards)

Proper English model card: Glint-protocol metrics, honest positioning, corrected corpus/positions

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub