Skip to content
conftrace_

toxicity score

4 papers

Explore in graph

Co-occurring keywords · papers mentioning both

jailbreak attack (2) vision-language model (2) adversarial attack (2) retention score (1) multimodal context (1) self correction (1) unsafe content (1) visual context injection (1) vision-centric jailbreak (1) custom finetuning (1)

Papers

ELITE: Enhanced Language-Image Toxicity Evaluation for Safety ICML 2025 Retention Score: Quantifying Jailbreak Risks for Vision Language Models AAAI 2025 Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection EMNLP 2025 Learning and Forgetting Unsafe Examples in Large Language Models ICML 2024