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

Researchers introduce TRACED, a framework that evaluates AI reasoning quality through geometric analysis rather than traditional scalar probabilities. The system identifies correct reasoning as high-progress stable trajectories, while AI hallucinations show low-progress unstable patterns with high curvature fluctuations.

• →TRACED framework uses geometric kinematics to assess LLM reasoning quality through Progress and Stability metrics.
• →Correct reasoning displays high-progress stable trajectories while hallucinations show stalled displacement with high curvature fluctuations.
• →The framework achieves competitive performance and superior robustness across diverse benchmarks.
• →High curvature maps to 'Hesitation Loops' and displacement to 'Certainty Accumulation' in machine reasoning.
• →This approach provides a physical lens to decode internal dynamics of AI thought processes.