Language models can explain neurons in language modelsby AI Generated Robotic Contentin FAANGon Posted on May 10, 2023We use GPT-4 to automatically write explanations for the behavior of neurons in large language models and to score those explanations. We release a dataset of these (imperfect) explanations and scores for every neuron in GPT-2.Share this article with your network:TwitterFacebookRedditLinkedInEmailLike this:Like Loading...