from Hacker News

COSP and USP: New methods to advance reasoning in LLMs

by TheIronYuppie on 1/25/24, 7:10 PM with 4 comments

• by sgt101 on 1/25/24, 7:29 PM

  Self adaptive prompting (prompt generation based on some sort of feedback from the model I think).

  USP: https://arxiv.org/abs/2305.14926

  COSP: https://aclanthology.org/2023.findings-acl.216/

  A hall of mirrors or a wonderful telescope? Time will tell.

• by TwentyPosts on 1/25/24, 7:23 PM

  This appears to be for "paid subscribers".