News for hackers. Stuff that matters.
Re:Task-Specific LLM Evals That Do and Don't Work by pizza on Monday December 09, 2024 @01:26PM (#42368912)
It seems to me that it would make sense to just include more <BOS>-like meta tokens at the beginning in such cases, and have them as a prefixed scratch space that can be suppressed by treating them as non-output tokens.

There are no comments on this comment yet.