Making LLMs Say What They Think: Measuring and Improving CoT-Interpretability Alignment
Paper • 2609.38972 • Published • 28
YihuaiHong
YihuaiHong
AI & ML interests
NLP, Interpretability, LLM
Recent Activity
authored a paper 4 days ago
Precise In-Parameter Concept Erasure in Large Language Models authored a paper 4 days ago
Dissecting Fine-Tuning Unlearning in Large Language Models authored a paper 4 days ago
The Reasoning-Memorization Interplay in Language Models Is Mediated by a
Single DirectionOrganizations
None yet