johnbean393/qwen3.5-4b-reasoning-effort-grpo
1
Hourly adapter snapshot at step 377
Hourly adapter snapshot at step 298
Hourly adapter snapshot at step 219
Hourly adapter snapshot at step 139
Hourly adapter snapshot at step 65 (recovered)
Add frozen 512-problem SGLang natural-length calibration
Add frozen 512-problem SGLang natural-length calibration
initial commit
