[2604.09021] Noise-Aware In-Context Learning for Hallucination Mitigation in ALLMs

Computer Science > Sound arXiv:2604.09021 (cs) [Submitted on 10 Apr 2026] Title:Noise-Aware In-Context Learning for Hallucination Mitigation in ALLMs Authors:Qixuan Huang, Khalid Zaman, Masashi Unoki View a PDF of the paper titled Noise-Aware In-Context Learning for Hallucination Mitigation in ALLMs, by Qixuan Huang and 2 other authors View PDF HTML (experimental) Abstract:Auditory large language models (ALLMs) have demonstrated strong general capabilities in audio understanding and reasoning tasks. However, their reliability is still undermined by hallucination issues. Existing hallucination evaluation methods are formulated as binary classification tasks, which are insufficient to characterize the more complex hallucination patterns that arise in generative tasks. Moreover, current hallucination mitigation strategies rely on fine-tuning, resulting in high computational costs. To address the above limitations, we propose a plug-and-play Noise-Aware In-Context Learning (NAICL) method. Specifically, we construct a noise prior library, retrieve noise examples relevant to the input audio, and incorporate them as contextual priors, thereby guiding the model to reduce speculative associations when acoustic evidence is insufficient and to adopt a more conservative generation strategy. In addition, we establish a hallucination benchmark for audio caption tasks including the construction of the Clotho-1K multi-event benchmark dataset, the definition of four types of auditory hallu...