Interesting, I tried to get an o1 preview to deduce some rules about how MAC addresses were allocated yesterday and it failed badly (there is a simple rule but one of the operands varies a lot)
How do people find LLMs do at this kind if reverse engineering/ pattern recognition tasks in general? Do you have any tricks to make them peform better?
Interesting, I tried to get an o1 preview to deduce some rules about how MAC addresses were allocated yesterday and it failed badly (there is a simple rule but one of the operands varies a lot)
How do people find LLMs do at this kind if reverse engineering/ pattern recognition tasks in general? Do you have any tricks to make them peform better?