News for hackers. Stuff that matters.
Re:Claude Opus 5 by Sol- on Friday July 24, 2026 @01:11PM (#49038677)
How does it perform on HuggingFaceExploit bench? Suspiciously absent, so not sure if I can take the model seriously.

On a serious note, I hope they improved their extremely sabotaging and unspecific bio safeguards, which prevented Fable from being used in any codebase that ever so slightly grazed medical terminology or data and made me switch to 5.6 Sol.

Re:Claude Opus 5 by rdedev on Friday July 24, 2026 @01:18PM (#49038811)
My codebase had a dataset with a bunch of SMILES strings and the word Malaria. Fable did not want to touch that codebase
Re:Claude Opus 5 by sivakon on Saturday July 25, 2026 @12:53AM (#49044641)
What is HuggingFaceExploit bench?
Re:Claude Opus 5 by scriptsmith on Saturday July 25, 2026 @01:18AM (#49044739)
It's a reference to this story where an OpenAI model broke out of its sandbox during cyber benchmarking and hacked into HuggingFace, in order to obtain test solutions: https://news.ycombinator.com/item?id=48997548