1 paper
Hiroki Furuta, Kuang-Huei Lee, Shixiang Shane Gu +4
Many algorithms for aligning LLMs with human preferences assume that human preferences are binary and deterministic. However, human preferences can vary across individuals, and the…