Cofca: A Step-Wise Counterfactual Multi-hop QA benchmark

Jian Wu, Linyi Yang, Zhen Wang +2

While Large Language Models (LLMs) excel in question-answering (QA) tasks, their real reasoning abilities on multiple evidence retrieval and integration on Multi-hop QA tasks remai…