NeurIPS paper: [MASK] tokens betray jailbreaks in masked diffusion language models
kastnerkyle · x · 2026-11-11
A NeurIPS paper shows that the [MASK] placeholders in a masked diffusion LLM's response are themselves a jailbreak signal: by modeling the probability distributions at masked positions for normal vs. jailbreak requests and thresholding a score s(x), some jailbreaks are easier to detect from the masks than from the prompt itself.
More from Safety
- BioVault launches a password manager with no master password, opens via face scan — bittingthembits · 2026-11-11
- Antitrust risk as a blocker to AI safety coordination: real concern or convenient excuse? — lxrjl · 2026-11-11
- Bluesky's PDS architecture makes its valuable social graph data a free-for-all — cephaloform · 2026-11-11
- After a $129M deepfake heist, World ID launches Zoom human verification — BecauseCulture · 2026-11-11
- Thinking Machines hires policy researchers to push open weights, decentralized AI — luke_drago_ · 2026-11-11
- Columbia law professor Neel Guha to discuss evaluating legal AI at Crosby Intelligence event — graceisford · 2026-11-11