[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