With coauthors from HLS and OpenAI, Manon Revel introduces evaluative metrics for reward models' alignment with values expressed in training datasets. "The importance of having a high-quality ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback