Remember: "reasoning models" aren't reasoning. They just incorporate additional reinforcement learning after brute force training data fitting.
It's just a different kind of #AlgorithmicMimicry
https://arxiv.org/abs/2307.07515
=> More informations about this toot | View the thread | More toots from yoginho@spore.social
=> View algorithmicmimicry tag This content has been proxied by September (3851b).Proxy Information
text/gemini