#artificial intelligence
May 2026
PAIR: Prefix-Aware Internal Reward Model for Multi-Turn Agent Optimization
The Prefix-Aware Internal Reward (PAIR), a two-stage model with a frozen hidden-state probe estimating belief-consistency and a lightweight attention-based head correcting it toward grounded correctness, is proposed, enabling dense step-level reward signals for GRPO training without external model calls, ground-truth d...
Wonjoong Kim, Yeonjun In, Sangwu Park et al.
· arXiv.org · 5 citations