AIKIT
Beyond Scalars: Evaluating and Understanding LLM Reasoning via Geometric Progress and Stability | AIKIT