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%

v15: ROUGE-L 0.960, 70% exact match — LR 2.5e-5 breakthrough