Skip to content
X · @teortaxesTex · X / Twitter

SIGNIFICANT endorsement of the EdgeBench "eval" (Dan's own evals heavily shaped research programs) Fable on why I'm correct as usual:

SIGNIFICANT endorsement of the EdgeBench "eval"(Dan's own evals heavily shaped research programs)Fable on why I'm correct as usual:Dan Hendrycks: @teortaxesTex It's a very substantial work.