Noise-Aware In-Context Learning for Hallucination Mitigation in ALLMs

ArXi:2604.09021v1 Announce Type: cross Auditory large language models (ALLMs) have nstrated 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 complex hallucination patterns that arise in generative tasks. Moreover, current hallucination mitigation strategies rely on fine-tuning, resulting in high computational costs.