ResearchKraken Tensor Core EM Side-Channel DNN Parameter ExtractionarXiv·high signalXBlueskyLinkedInCopy linkFirst DNN parameter extraction from GPU Tensor Cores via EM side-channel. Proves inference hardware is attack surface.SourceSource pagearXiv↳ Follow the threadStack layer / ContrastOdin Runs All 32 Llama-3-8B Transformer Layers Under FHE in 366 Seconds on One H100, 4.51x Faster Than THORarXiv 2609.12378Policy dependency / Stack layerRIPPLE: an edit confined to one prompt-policy segment changes downstream behavior, so replay candidate edits after previously accepted ones before persistingarXiv 2609.12127Stack layerPermutation-Based Model Confidentiality in Hybrid FHE Breaks in d+1 Queries Per LayerarXiv 2609.12911Stack layerA seven-person team trained open-weight cyber agents to 63.24% on CyberGym, rank 10 overallarXiv / HuggingFace Daily PapersStack layerA Jailbreak SoK Finds Low Final-Response Attack Success Hides Compromised Planning, Memory, and Tool StatearXiv 2609.12413Stack layer / Threat patternUnlearning Methods That Pass TOFU and MUSE Still Leak the Secret on 22-86% of Queries Once the Model Is an AgentarXiv 2609.12808Policy dependency / Stack layerSGLang Hit With Unauthenticated Pickle RCE via /update_weights_from_tensor, the Fourth Critical Inference-Stack CVE in Four WeeksCERT Coordination CenterStack layerAgent-Integrated Software names the coordination problem of users editing shared objects while delegated work runsarXiv