Towards Faithful Natural Language Explanations: A Study Using Activation Patching in Large Language Models.

Wei Jie Yeo, Ranjan Satapathy, Erik Cambria

Browse the full EMNLP paper archive.