pith. machine review for the scientific record. sign in

BadCLIP: Trigger- Aware Prompt Learning for Backdoor Attacks on CLIP

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it

fields

cs.CR 1

years

2026 1

verdicts

UNVERDICTED 1

representative citing papers

Cross-Modal Backdoors in Multimodal Large Language Models

cs.CR · 2026-05-08 · unverdicted · novelty 8.0

Poisoning a single connector in MLLMs establishes a reusable latent backdoor pathway that transfers across modalities with over 95% attack success rate under bounded perturbations.

citing papers explorer

Showing 1 of 1 citing paper.

  • Cross-Modal Backdoors in Multimodal Large Language Models cs.CR · 2026-05-08 · unverdicted · none · ref 13

    Poisoning a single connector in MLLMs establishes a reusable latent backdoor pathway that transfers across modalities with over 95% attack success rate under bounded perturbations.