Commit History

rl: pin TRITON_ATTN backend (FLASHINFER silently breaks vllm-lens injection on Blackwell)
28ad77d

ceselder Claude Fable 5 commited on

rl: true per-token entropy bonus (maximize r + beta*H, explicit diversity knob, no KL)
73adae4

ceselder Claude Fable 5 commited on

maxact-fast: hardening fixes (6 bugs) + Dr.GRPO vllm-lens rl.py + sbatch; Probe = universal direction
48db85f

ceselder Claude Fable 5 commited on