1Do LLMs Reason, or Do They Just Predict Math Text? (opens in new tab)(daridor.blog)4beagle31mo ago2Save
2AST-filtered eval() is not a sandbox: Severity 10 CVE-2026-26030, and others (opens in new tab)(daridor.blog)2beagle33mo ago0Save
3When Poetry Meets AI Safety: A Critical Look at "Universal" Jailbreaks (opens in new tab)(daridor.blog)2beagle37mo ago0Save
