Emergent misalignment (EM), in which fine-tuning on a narrow task induces misaligned behaviour on unrelated prompts, has been demonstrated across a wide range of datasets, model families, and training setups. Yet, most studies continue to rely on the evaluation approach introduced in the original EM paper, producing ra...
A. Domingues, Mubashara Akhtar, Bhada Yun et al.· 0 citations
The findings suggest that multimodal EM reflects a behavioral shift rather than a general loss of capability, extending beyond text to the visual modality.
Shun-Chang Liu, Lukas Fluri, Xin Chen et al.· 0 citations
This paper introduces CryptanalysisBench, 191 tasks across six families of cryptographic primitives drawn primarily from four NIST standardization competitions, and releases CryptanalysisBench as a tool to help track if (or when) AI cryptanalysis becomes a serious factor and as a scaffold for stress-testing candidate s...
Lukas Fluri, Avital Shafran, Nicholas Carlini et al.· arXiv.org· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.