Promote Phase 3A LoRA — Qwen 3B beats heuristic on HARD, 100% rogue catch 90452ca helloAK96 Claude Opus 4.7 commited on 14 days ago
Promote Phase 2 LoRA (curriculum + LR=2e-5 + r=32) as the live trained lane f89a0e8 helloAK96 Claude Opus 4.7 commited on 14 days ago
Ship trained-policy artifact: evaluation_summary.txt 83ffa3f verified helloAK96 commited on 15 days ago
helloAK96 Claude Opus 4.7 commited on 