causal
Pre-Training Lacks Novel Discovery Skill
The pre-training objective gives models flexible general knowledge but does not imbue the skill of making novel connections or doing research; significant RL on research-like tasks is likely needed before models can approach making novel scientific discoveries.
causalpending
Speaker
Sholto DouglasEvidence Quote
“I think at a minimum you need significant RL in at least similar things to be able to approach making novel discoveries.”
Created: 6/13/2026, 3:36:46 AM
Your notes
Loading notes...