NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
22 results for “keyword:"multimodal safety"” in ICLR 2027. 11 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]The Perceptual Bandwidth Tax: Matched Evaluation of Emotion and Accent as Jailbreak Surfaces in End-to-End Audio-Language Models▲452.75
[PDF]The Refraction Gap: How Optical Context Compression Inadvertently Bypasses Safety Alignment▲3.0k2.33
[PDF]Jailbreaking Multimodal Large Language Models through Generative Visual Semantic Ciphers▲19k2.50
[PDF]Hyperbolic Detoxification for Vision-Language Model Safety via Intent-Risk Calibration▼5.8k2.33
[PDF]SEFIR: Safety Detection in Frozen Multimodal Large Language Models via Internal Evidence Fusion▲3972.67
[PDF]Seeing Safety Across Languages: Disentangling Multilingual Safety from Visual-Language Failures in Multimodal Large Language Models▼1361.33
Withdrawn or desk rejected (11)
[PDF]Judgement Day: An Anatomy of Successful Multimodal Attacks on Safety-Critical AI SystemsDesk rejected
[PDF]UnifiedAttack: Evaluating the Safety of Multimodal Large Language Models in Synergistic Harmful Image-Text GenerationDesk rejected
[PDF]Non-Intrusive Safety Alignment for Vision-Language Models via Automated Output RewritingWithdrawn1.67
[PDF]Safe Vision-Language Models via Sparse Autoencoder Feature-Guided Embedding SteeringWithdrawn1.75
[PDF]Where and How to Steer Representations for Jailbreak Attacks on Multimodal Large Language ModelsWithdrawn2.00
[PDF]VPA-Guard: Defending and Benchmarking Image-to-Video Generation Against Visual Prompt AttacksWithdrawn1.67