TRIAGE: Direction-Aware Mismatch Stabilization of Native NVFP4 Reinforcement Learning

Excerpt

arXiv:2610.07043v1 Announce Type: new Abstract: Low-precision execution can substantially accelerate reinforcement learning (RL) for large language models, but discrepancies between learner and sampler execution can destabilize policy optimization. In this paper, we characterize the interaction between mismatch and the policy-gradient direction, distinguishing locally amplifying from contracting update contributions that mismatch magnitude alone cannot identify. In native NVFP4 runs, we observe