Stable Rl With Llms

Justifies token-level optimization objectives as first-order approximation to sequence rewards via Routing Replay, which fixes MoE routers during policy optimization to reduce training-inference discrepancy. Use when scaling RL training on large MoE language models.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit 4654a76d5f

Frequently asked questions

npx get-skill add adu2021/stable-rl-with-llms