Why aren't LLM's trained on their own Chain Of Thought?

3 points by simianwords ↗ HN
I realise that without allowing reasoning tokens, a model performs very poorly. It can't perform simple arithmetic or simple logic and hallucinates a bit.

But by allowing it to think a bit and then answer, the result is much better and way more trustable.

This shows a clean RL environment.. or just a nice data-set. Where you prompt the model two times - one without allowing thinking and one with thinking. Penalise the result from non thinking if the result contradicts the answer obtained from thinking.

1 comment

[ 2.1 ms ] story [ 11.3 ms ] thread
For the same reason that anyone's reasoning process and answers to random exam questions are never used as textbooks: if the reasoning is not guaranteed to be right, why would you want to make that training material?