Procedural Knowledge in Pretraining Drives Reasoning in Large Language Models

Shared on Bluesky by 2 AI experts

• Tim Kellogg @timkellogg.me: this is one of my favorite papers, similar method. It convinced me a that LLMs weren’t merely autocompleting they showed that EVEN WHEN the… →
• John David Pressman @jdp.extropian.net amplified

  Tim Kellogg @timkellogg.me

  this is one of my favorite papers, similar method. It convinced me a that LLMs weren’t merely autocompleting they showed that EVEN WHEN the answer to a math problem was present in the dataset, the agent didn’t reference…

  View on Bluesky →