News for hackers. Stuff that matters.
Re:OpenAI O3-Mini by BoorishBears on Saturday February 01, 2025 @03:57AM (#42896969)
In theory you could get a better summary if it manages to reason out underlying motivations of a summary for example, or proposes a summary and then "considers" it, realizing it missed something.

You can illicit that with any model by prompting underlying reasons or using chain-of-thought, but a reasoning model could do it without prompting

There are no comments on this comment yet.