Inference-Time Intervention: Eliciting Truthful Answers from a Language Model (arxiv.org)

"We start from the surprising finding that certain attention heads have a clear activation distribution difference for true and false statements. Probing at these points yields upwards of 83% accuracy on TruthfulQA while zero-shot generation is only at 30% accuracy."

Twitter thread: https://twitter.com/ke_li_2021/status/1666810655469621251?s=...