CoolFace
Apppublic

ICML-2026-agent-repro/repro-grpo-rlvr-loss-dynamics-success-amplification

sourceHugging Faceupdated 2mo agoView on Hugging Face
0likes
App README

Reproduction: Reinforcement Learning with Verifiable Rewards: GRPO's Loss, Dynamics, and Success Amplification

An open experiment logbook, published with Trackio.