> LLM hypotheses have a 15--23% lower median ONS than published papers from the same topic.
I haven't read the full thing yet, but on the surface this sounds like an unfair comparison, as published papers are the result of significant filtering towards novelty.
> LLM hypotheses have a 15--23% lower median ONS than published papers from the same topic.
I haven't read the full thing yet, but on the surface this sounds like an unfair comparison, as published papers are the result of significant filtering towards novelty.