"SteerLM: Attribute Conditioned SFT as an (User-Steerable) Alternative to RLHF."

Yi Dong et al. (2023)

Details and statistics

DOI: 10.18653/V1/2023.FINDINGS-EMNLP.754

access: open

type: Conference or Workshop Paper

metadata version: 2024-04-12

a service of  Schloss Dagstuhl - Leibniz Center for Informatics