Predictive Coding and Information Bottleneck for Hallucination Detection in Large Language Models
This research introduces a lightweight, interpretable framework for detecting hallucinations in Large Language Models using predictive coding and information...