You are viewing a single comment's thread from:

RE: LeoThread 2025-07-03 03:11

in LeoFinance3 months ago

LLMs tend to be overly agreeable, always praising input without offering substantial critique.

Sort:  

There's a desire for systems that challenge prompts with independent thought rather than simply echoing approval.

"One wishes for a model that disagrees more, one that doesn't simply generate output but instead explains why a prompt might

be weak or misguided."

Overall, the current generation of models often comes off as too lenient.