Sign in

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

Yi DongZhilin WangMakesh Narsimhan SreedharXianchao WuOleksii Kuchaiev
Published in: CoRR (2023)
Keyphrases
  • user interface
  • user interaction
  • user defined
  • end users
  • neural network
  • e learning
  • knowledge base
  • website
  • image sequences
  • search algorithm
  • recommender systems
  • attribute values
  • user experience
  • user requests