#ICML 2024 How do large language models (LLM) reach their decisions? Our latest research project, SelfIE, is the first to use an LLM to explain the same LLM's internals. The interpretation can be used for safety alignment and understanding hallucinations. selfie.cs.columbia.edu

🚨New #ACL2025 paper! Today’s “safe” language models can look unbiased—but alignment can actually make them more biased implicitly by reducing their sensitivity to race-related associations. 🧵Find out more below!