Sign up
Sign in
…nt, the chosen and rejected responses are scored by both the trained and the frozen language model. This score is the product of the probabilities associated with the desired response token for each step.
João Lages
Masum Hasan
Follow
--
Share
Is this the language modeling score?
Researcher in NLP and Machine Learning | masumhasan.com
Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams
