#7B

The reward-shaping trick means theory of mind gets baked into the weights rather than bolted on at inference, so negotiation and cooperation gains come with zero added serving cost. Agent builders working on multi-agent negotiation or role-play can borrow this training recipe directly. Compared with the usual memory-summary heuristics, an explicit partner-model reward targets the exact failure mode of misreading intent.

rohanpaul_ai@rohanpaul_aiSource

For developers, this is a rare look at the full process of training a large model from scratch, with all resources open for anyone to use or study.

量子位@思邈Source