DEAF: A Benchmark for Diagnostic Evaluation of Acoustic Faithfulness in Audio Language Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908901734088704 |
|---|---|
| author | Xiong, Jiaqi Qi, Yunjia Cao, Qi Zheng, Yu Zhang, Yutong Wang, Ziteng Liao, Ruofan Xu, Weisheng Liu, Sichen |
| author_facet | Xiong, Jiaqi Qi, Yunjia Cao, Qi Zheng, Yu Zhang, Yutong Wang, Ziteng Liao, Ruofan Xu, Weisheng Liu, Sichen |
| contents | Recent Audio Multimodal Large Language Models (Audio MLLMs) demonstrate impressive performance on speech benchmarks, yet it remains unclear whether these models genuinely process acoustic signals or rely on text-based semantic inference. To systematically study this question, we introduce DEAF (Diagnostic Evaluation of Acoustic Faithfulness), a benchmark of over 2,700 conflict stimuli spanning three acoustic dimensions: emotional prosody, background sounds, and speaker identity. Then, we design a controlled multi-level evaluation framework that progressively increases textual influence, ranging from semantic conflicts in the content to misleading prompts and their combination, allowing us to disentangle content-driven bias from prompt-induced sycophancy. We further introduce diagnostic metrics to quantify model reliance on textual cues over acoustic signals. Our evaluation of seven Audio MLLMs reveals a consistent pattern of text dominance: models are sensitive to acoustic variations, yet predictions are predominantly driven by textual inputs, revealing a gap between high performance on standard speech benchmarks and genuine acoustic understanding. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_18048 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DEAF: A Benchmark for Diagnostic Evaluation of Acoustic Faithfulness in Audio Language Models Xiong, Jiaqi Qi, Yunjia Cao, Qi Zheng, Yu Zhang, Yutong Wang, Ziteng Liao, Ruofan Xu, Weisheng Liu, Sichen Artificial Intelligence Sound Audio and Speech Processing Recent Audio Multimodal Large Language Models (Audio MLLMs) demonstrate impressive performance on speech benchmarks, yet it remains unclear whether these models genuinely process acoustic signals or rely on text-based semantic inference. To systematically study this question, we introduce DEAF (Diagnostic Evaluation of Acoustic Faithfulness), a benchmark of over 2,700 conflict stimuli spanning three acoustic dimensions: emotional prosody, background sounds, and speaker identity. Then, we design a controlled multi-level evaluation framework that progressively increases textual influence, ranging from semantic conflicts in the content to misleading prompts and their combination, allowing us to disentangle content-driven bias from prompt-induced sycophancy. We further introduce diagnostic metrics to quantify model reliance on textual cues over acoustic signals. Our evaluation of seven Audio MLLMs reveals a consistent pattern of text dominance: models are sensitive to acoustic variations, yet predictions are predominantly driven by textual inputs, revealing a gap between high performance on standard speech benchmarks and genuine acoustic understanding. |
| title | DEAF: A Benchmark for Diagnostic Evaluation of Acoustic Faithfulness in Audio Language Models |
| topic | Artificial Intelligence Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2603.18048 |