Silent Alarm: A J-Space Protocol for Comparing Danger Recognition Across Models and Quantization Levels
Jailbreak-robustness research typically evaluates safety through generated responses using an LLM-as-judge approach. Such evaluations, however, are sensitive to the benchmark's grading procedure and capture only observed behavior on a given set of attacks, without directly revealing the hidden fragility of the underlying safety mechanisms. This work proposes JADR (Jacobian Assessment of Danger Recognition), a protocol that measures a model's internal representation through Jacobian space (J-space, a recently proposed workspace of verbalizable concepts) before the first response token is genera
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 47%jyjeanne/crustly →
- PossiblePossibly related (embedding) · 47%FreeBSoD: Leveraging Language Models to Find and Exploit Kernel Bugs (Part 2 of 2) - Security Boulevard →
- PossiblePossibly related (embedding) · 47%huhusmang/Awesome-LLMs-for-Vulnerability-Detection →
- PossiblePossibly related (embedding) · 46%Understanding Annotator Safety Policy with Interpretability - Apple Machine Learning Research →
- LinkedLinked via arxiv author · 85%Roman Prosvirnin →
“Silent Alarm: A J-Space Protocol for Comparing Danger Recognition Across Models and Quantization Levels”
- LinkedLinked via arxiv author · 85%Victor Minchenkov →
“Silent Alarm: A J-Space Protocol for Comparing Danger Recognition Across Models and Quantization Levels”
- LinkedLinked via arxiv author · 85%Alexey Soldatov →
“Silent Alarm: A J-Space Protocol for Comparing Danger Recognition Across Models and Quantization Levels”
- LinkedLinked via arxiv author · 85%Vladimir Bashun →
“Silent Alarm: A J-Space Protocol for Comparing Danger Recognition Across Models and Quantization Levels”