Critic Training Details and Stability
ziv_ravid · x · 2026-07-11
This part details the implementation of training a critic:
- To make the critic trainable, the authors froze the attention layers and only updated the MoE projections, as gradient explosions primarily originate from attention.
- The critic is updated twice per policy step.
- For multi-turn trajectories, they skip observation tokens in the GAE.
Overall, it explains how to handle numerical instabilities in value models and trajectories during agent/RL training.
Related event: GLM Team Proposes SAO Algorithm for Asynchronous Agent RL(15 posts)→
More from Research
- Structural ensembles beat single predictions in TCR:pMHC generalization study — quaidmorris · 2026-07-22
- RSS launches under OMSF to push structural biology data modeling at scale — MoAlQuraishi · 2026-07-22
- enFoldX tops 8 neoantigen scans and an unseen-peptide benchmark — quaidmorris · 2026-07-22
- enFoldX reaches AUC 0.82 on human VDJdb and transfers to mouse at 0.76 — quaidmorris · 2026-07-22
- enFoldX gains accuracy as AF3 ensemble disagreement rises for non-binders — quaidmorris · 2026-07-22
- A 3D ray plot shows how hard this Jacobian counterexample is to read — moultano · 2026-07-22