When Good Sounds Go Adversarial: Jailbreaking Audio-Language Models with Benign Inputs
Paper
•
2508.03365
•
Published
•
5
Hisakias Dingeto
hisku
·
AI & ML interests
NLP, Meta-Learning
Recent Activity
liked
a model
1 day ago
anton-hugging/TimeOmni-1-7B
authored
a paper
6 months ago
When Good Sounds Go Adversarial: Jailbreaking Audio-Language Models with
Benign Inputs
authored
a paper
6 months ago
Eliciting and Analyzing Emergent Misalignment in State-of-the-Art Large
Language Models
Organizations
authored
2
papers
6 months ago