@georgejrjrjr @lateinteraction I generally agree, with the caveat that the most capable models shouldn't need highly tuned prompts, so it really depends on what you're trying to measure. But for…

Posters: xlr8harder. xlr8harder

2 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/xlr8harder/status/1750032377785786447