Understanding In-Context Multimodal Jailbreaks via Posterior Reweighting
Updated 2 h ago · first seen 11 Sept 2026
paper_01M294FPR93HPSR7MWMEMEE6ZM
- Published
- 11 Sept 2026
- T1 · 2 h ago
- arXiv
- 2609.10613
- T1 · 2 h ago
- Category
- cs.CR
- T1 · 2 h ago
Abstract
In-context learning (ICL) jailbreaks reveal a critical vulnerability in multimodal large language models (MLLMs): harmful demonstrations in the prompt can induce unsafe outputs without modifying model parameters. Despite extensive empirical evidence, existing work lacks a principled understanding of why such jailbreaks reliably succeed or how their effectiveness scales with context composition. We propose a posterior reweighting framework that models a safety-aligned MLLM as implicitly operating over competing behavioral modes, and interprets in-context demonstrations as inference-time evidence that dynamically shifts the model's posterior preference between safe and harmful behaviors. This view formalizes jailbreak as a process of evidence accumulation, yielding predictive scaling laws with respect to demonstration count, harmful ratio, adversarial strength, and semantic diversity. Guided by this framework, we introduce a posterior-aware inference-time defense that adaptively injects benign counter-evidence based on estimated risk, effectively suppressing harmful posterior drift while preserving model utility. Compared to existing in-context defenses, our method achieves a significantly improved robustness-utility trade-off under a fixed intervention budget. Together, our results establish posterior reweighting as a unifying and predictive framework for understanding and mitigating ICL jailbreak in MLLMs.
Authors 4
Xu Zhang, Dev Mistry, Xiang Xu, Ren Wang
Specification
- Official page
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Arxiv announce type
- cross
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- arXiv id
- 2609.10613
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Categories
- cs.CR, cs.LG
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Primary category
- cs.CR
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
- Published
- 11 Sept 2026
Source:arXiv (Atom API + RSS)T1observed 2 h agohigh
Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →
Provenance
Attributed facts
9
Source tiers
T19
Freshest observation
2 h ago
Conflicts
None
No models linked to this paper yet.
- Authors
- Xu Zhang, Dev Mistry, Xiang Xu
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history · Abstract
Abstractabstract1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| In-context learning (ICL) jailbreaks reveal a critical vulnerability in multimodal large language models (MLLMs): harmful demonstrations in the prompt can induce unsafe outputs without modifying model parameters. Despite extensive empirical evidence, existing work lacks a principled understanding of why such jailbreaks reliably succeed or how their effectiveness scales with context composition. We propose a posterior reweighting framework that models a safety-aligned MLLM as implicitly operating over competing behavioral modes, and interprets in-context demonstrations as inference-time evidence that dynamically shifts the model's posterior preference between safe and harmful behaviors. This view formalizes jailbreak as a process of evidence accumulation, yielding predictive scaling laws with respect to demonstration count, harmful ratio, adversarial strength, and semantic diversity. Guided by this framework, we introduce a posterior-aware inference-time defense that adaptively injects benign counter-evidence based on estimated risk, effectively suppressing harmful posterior drift while preserving model utility. Compared to existing in-context defenses, our method achieves a significantly improved robustness-utility trade-off under a fixed intervention budget. Together, our results establish posterior reweighting as a unifying and predictive framework for understanding and mitigating ICL jailbreak in MLLMs. | → current | current | arXiv (Atom API + RSS)T1 | high | deterministic |
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
New paper: Understanding In-Context Multimodal Jailbreaks via Posterior Reweighting
arxiv
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| arXiv (Atom API + RSS) | rss.arxiv.org/rss/cs.LG | feed | T1· Official | 14 min ago | 1 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.