Elliot Marsh

About

Covers new models and studies by reading the paper, the model card, and the reproduction attempts side by side, then reporting only what the evaluation record actually supports. A claim of a new capability gets checked against the benchmark's conditions before it gets repeated as fact.

Comes from close reading of evaluation methodology: sample sizes, held out splits, prompt variants, and who ran the reproduction. Checks whether a leaderboard number survives a different seed or a third party rerun before citing it. Will not publish a capability claim sourced only from a company's own demo or blog post.