Perplexity says it attacked its SPACE sandbox with frontier models. In Escaping SPACE: Part I, nine model configurations with root inside the VM made 108 escape attempts, and none reached the host. The full research is linked in the post.
Part of a thread · earlier post
We tested SPACE by attacking it with frontier models. In Escaping SPACE: Part I, nine model configurations with root inside the VM made 108 escape attempts, and none reached the host.
Read the research: https://www.perplexity.ai/hub/blog/escaping-space-part-i
Source: Perplexity · x.comPublished
