Learning to Reason with LLMs

Re:Learning to Reason with LLMs by dragonwriter on Thursday September 12, 2024 @01:59PM (#41523643)

Yeah, exposed chain of thought is more useful as a user, as well as being useful for training purposes.

[ Parent | on HN ]

Re:Learning to Reason with LLMs by riku_iki on Thursday September 12, 2024 @02:05PM (#41523721)

I think we may discover that model do some cryptic mess inside instead of some clean reasoning.

[ Parent | on HN ]

Re:Learning to Reason with LLMs by hadlock on Thursday September 12, 2024 @02:41PM (#41524118)

Loopback to: "my code works. why does my code work?"

[ Parent | on HN ]

Re:Learning to Reason with LLMs by MVissers on Thursday September 12, 2024 @05:06PM (#41525573)

I’d say depends. If the model iterates 100x I’d just say give me the output.

Same with problem solving in my brain: Sure, sometimes it helps to think out loud. But taking a break and let my unconcious do the work is helpful as well. For complex problems that’s actually nice.

I think eventually we don’t care as long as it works or we can easily debug it.

[ Parent | on HN ]