Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2603.29263 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914433716977664 |
|---|---|
| author | Seth, Ashish Kumar, Sonal Selvakumar, Ramaneswaran Anand, Nishit Tyagi, Utkarsh Seetharaman, Prem Duraiswami, Ramani Manocha, Dinesh |
| author_facet | Seth, Ashish Kumar, Sonal Selvakumar, Ramaneswaran Anand, Nishit Tyagi, Utkarsh Seetharaman, Prem Duraiswami, Ramani Manocha, Dinesh |
| contents | Large Audio Language Models (LALMs) achieve strong performance on audio-language tasks; however, their reliability in real-world settings remains underexplored. We introduce Audio Hallucination Attacks (AHA), an attack suite called AHA-Eval, comprising 6.5K QA pairs designed to test whether LALMs genuinely ground their responses in the audio input. AHA targets two attack surfaces: (i) query-based attacks, which exploit question structure to induce hallucinations about absent sounds, and (ii) audio-based attacks, which inject synthetic speech describing non-existent events into the audio stream. Evaluating state-of-the-art LALMs, including Audio Flamingo 3 and Gemini 3 Pro, we observe high attack success rates of 95.35% and 79.65%, respectively, revealing a reliability gap that is hidden by standard benchmark performance. To mitigate this, we propose a 120K QA post-alignment dataset, AHA-Guard, which successfully reduces attack success rates by up to 49%. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_29263 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Audio Hallucination Attacks: Probing the Reliability of Large Audio Language Models Seth, Ashish Kumar, Sonal Selvakumar, Ramaneswaran Anand, Nishit Tyagi, Utkarsh Seetharaman, Prem Duraiswami, Ramani Manocha, Dinesh Sound Large Audio Language Models (LALMs) achieve strong performance on audio-language tasks; however, their reliability in real-world settings remains underexplored. We introduce Audio Hallucination Attacks (AHA), an attack suite called AHA-Eval, comprising 6.5K QA pairs designed to test whether LALMs genuinely ground their responses in the audio input. AHA targets two attack surfaces: (i) query-based attacks, which exploit question structure to induce hallucinations about absent sounds, and (ii) audio-based attacks, which inject synthetic speech describing non-existent events into the audio stream. Evaluating state-of-the-art LALMs, including Audio Flamingo 3 and Gemini 3 Pro, we observe high attack success rates of 95.35% and 79.65%, respectively, revealing a reliability gap that is hidden by standard benchmark performance. To mitigate this, we propose a 120K QA post-alignment dataset, AHA-Guard, which successfully reduces attack success rates by up to 49%. |
| title | Audio Hallucination Attacks: Probing the Reliability of Large Audio Language Models |
| topic | Sound |
| url | https://arxiv.org/abs/2603.29263 |