3 ms·

Can you provide more details? I'm curious about the performance and limitations of these models.

• spacecadet 2y ago
  Like all LLMs you greatly benefit from prior experience or you risk just falling for hallucinations which is a limitation of a non-deterministic black box, and degrades performance relative to the task. Ive commented in other threads, LLMs are great at amplifying my output in an area I already have domain knowledge in. I think this is why people fail to realize any gains or give up, they think it will unlock areas they dont fully understand themselves. Blind leading the blind problem.