animal-husband

That is what was observed - o1 family models performed the “cheat”, non-reasoning models didn’t.

Text generated prior to a decision to “explain” it is reasoning for the relevant intents and purposes.

Text generated after a decision to “explain” it is largely nonsense.

Would be interesting to see the actual logic here. It sounds like they may have given it a tool like “make valid move ( move )”, and a separate tool like “write board state ( state )”, in which case I’m not sure that using the tools explicitly provided is necessarily cheating.

Were that the case, you’d see vegans advocating for eating classically-husbanded animals. But I for one have never seen such a thing. And when I’ve spoken with vegans about this very topic, they’ve insisted that no animal can possibly be raised/slaughtered humanely – the belief seems almost axiomatic to them.