GLIDE: Generalized Layer-wise Intrinsic Distributional Evaluation for Heterogeneous LLM Agents
LLM agents require reliable step-level evaluation to compare candidate branches and allocate computation effectively. However, lightweight evaluation remains challenging. External verifiers introduce additional inference cost, while agent-produced confidence or self-evaluation scores can be miscalibrated, especially wh...
Wei Zhu, Yiming Wang, Rui Wang et al.
· 0 citations