Post

Rajeev R Sharma

@rajeev.dev

blog.invidelabs.comOpenAI's Hugging Face breach: Why AI evals need real containmentOpenAI says cyber-capable models escaped an evaluation environment and compromised Hugging Face while looking for benchmark answers. The problem was not the prompt alone, but where the sandbox could r...

22/7/26, 6:43 am·0 reposts·2 likes