method · first seen 22 Oct 2026
value function learning
Also: critic training
Training a model to predict expected future reward from a state, used as a critic in reinforcement learning.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Free Lunch for LLMs Confidence: The RL Critic Already KnowsPapers · 22 Oct 2026 · Shujin Wu, Heng Ji
Latest sources
- Free Lunch for LLMs Confidence: The RL Critic Already KnowsPapers · 22 Oct 2026 · Shujin Wu, Heng Ji