2 ms·
What’s missing in that analogy is that humans tend to have a good hunch about when they have to think more and when they are “done”. LLMs seem to be missing a m
by layer8 2y ago
What’s missing in that analogy is that humans tend to have a good hunch about when they have to think more and when they are “done”. LLMs seem to be missing a mechanism for that kind of awareness.
- nico 2y agoGreat observation. Maybe an additional “routing model” could be trained to predict when it’s better to think more vs just using the current result
- sanxiyn 2y agoLLMs actually do have such hunch, they just don't utilize it. You can literally ask them "Would you do better if you started over?" and start over if answer is yes. This works. https://arxiv.org/abs/2410.02725 https://arxiv.org/abs/2410.02725