CounterBench: Evaluating and Improving Counterfactual Reasoning in Large Language Models
arXiv:2502.11008v2 Announce Type: replace
Abstract: Counterfactual reasoning is widely recognized as one of the most challenging and intricate aspects of causality in artificial intelligence. In this paper, we evaluate the performance of large languag…