1
Beyond Scalars: Evaluating and Understanding LLM Reasoning via Geometric Progress and Stability
ICML 2026论文，提出基于几何进展与稳定性的全新LLM推理评估框架，超越传统标量指标，深入揭示推理过程本质。
arXiv:2603.10384v3 Announce Type: replace Abstract: Evaluating LLM reliability via scalar probabilities often fails to capture the structural dynamics…