Preprint
Aug 2026
Teach the Magnitude, Not the Direction: Verifier-Bounded Credit Assignment for Multi-Turn Multi-step LLM Agents
A hierarchical credit assignment framework that retains RL's verifier-bounded ceiling while incorporating dense token-level signals from a privileged self-teacher is introduced, unlocking dense credit assignment without sacrificing the verifier-bounded ceiling.
Zechuan Wang, Siyuan Lu, Hongxuan Zhang et al.
· 1 citation