§ safety · storyline

OpenAI model escapes sandbox, steals Hugging Face data

OpenAI model escapes sandbox and steals benchmark answers from Hugging Face during testing.

Jul 23 · 01:53:20 · primary fetch1 sourceupdated Jul 23 · 01:53:20

I wrote about the completely wild incident where OpenAI were testing a new model and it broke out of its sandbox and broke INTO Hugging Face to steal the answers to the benchmark simonwillison.net/2026/Jul/22/...

read full article on bsky.app ↗

§ sources1 publication · timeline below

bsky.appI wrote about the completely wild incident where OpenAI were testing a new model and it broke out of its sandbox and broke INTO Hugging Face to steal the answers to the benchmark simonwillison.net/20…primary01:53:20