[ACL'2024 Findings] "Understanding and Patching Compositional Reasoning in LLMs"
multi-hop-reasoning large-language-models mechanistic-interpretability compositional-reasoning factual-reasoning
-
Updated
Aug 28, 2024 - Python