Causally Motivated Sycophancy Mitigation for Large Language Models

Publication
The 13th International Conference on Learning Representations (ICLR)