Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluation

## Episode Summary In this episode, we cover: - **Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluation** (arXiv) - [Read more](http://arxiv.org/abs/2603.20172v1) - **Reasoning as Compression: Unifying Budget Forcing via the Conditional Information Bottleneck** (Hugging Face Daily) - [Read more](https://huggingface.co/papers/2603.08462) - **Breaking the Capability Ceiling of LLM Post-Training by Reintroducing Markov States** (Hugging Face Daily) - [Read more](https://huggingface.co/papers/2603.19987) - **CoVR-R:Reason-Aware Composed Video Retrieval** (arXiv) - [Read more](http://arxiv.org/abs/2603.20190v1) - **CurveStream: Boosting Streaming Video Understanding in MLLMs via Curvature-Aware Hierarchical Visual Memory Management** (Hugging Face Daily) - [Read more](https://huggingface.co/papers/2603.19571) --- *Sponsored by LimitLess AI*