v51: composite=88.68 — see model card for benchmark deltas vs v45

v45: SFT+chained GRPO with ITN — 95.9% number accuracy, 97.0% filler-free, deletion behavior matches v36

v36: full-FT GRPO with substantive-deletion-aware reward — filler-free 96.9%, sub-del-15-long 0.64%

Full FT: ROUGE-L 0.907 — new record, +1.6 over prompted 2B

GRPO model: ROUGE-L 0.891 — matches prompted 2B

Upload folder using huggingface_hub