Trivial Prompt Reframing Bypasses Safety Guardrails in Google\'s MedGemma-4B
arXiv 2607.09804•12571ae55a8724798856a97a89fe2fbcb39c6c9688dd393c44ad6502a3600bc5
EDRHMACJWTadversarial-mlasr-attacksattack-surfacechaos-theorycryptographyecdlogelliptic-curveendpoint-securityimage-encryptionmachine-learningmedical-llmmodel-safetyprompt-jailbreakquantum-cryptographyreservoir-computingsocial-engineeringspeech-adversaryvishingvoice-phishingvoice-synthesiswebsocket
Paper metadata
- arXiv ID
- 2607.09804
- Version
- Not specified by this published record
- Category
- Computer Science — Cryptography and Security (cs.CR)
The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.
Evidence and limitations
- Source ID
- arxiv_cs_cr
- Record identifier
- 12571ae55a8724798856a97a89fe2fbcb39c6c9688dd393c44ad6502a3600bc5
- Enrichment time
- 2026-07-14T07:23:52Z
- AI-assisted enrichment
- Yes
This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.