4 ms·
That's unfortunate. When an LLM makes a mistake it's very helpful to read the CoT and see what went wrong (input error/instruction error/random shit)
by msp26 2y ago
That's unfortunate. When an LLM makes a mistake it's very helpful to read the CoT and see what went wrong (input error/instruction error/random shit)
- dragonwriter 2y agoYeah, exposed chain of thought is more useful as a user, as well as being useful for training purposes.
- riku_iki 2y agoI think we may discover that model do some cryptic mess inside instead of some clean reasoning.
- hadlock 2y agoLoopback to: "my code works. why does my code work?"
- MVissers 2y agoI’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.