Music Hallucination in Audio-Language Models: A Hierarchical Formulation and Empirical Study
This paper formalizes and empirically studies hallucination in audio-language models for music, a reliability issue relevant to AI safety in generative audio systems.
This paper formalizes and empirically studies hallucination in audio-language models for music, a reliability issue relevant to AI safety in generative audio systems.
Auto-scored: Academic study of model hallucination is a minor framing-level risk with no direct misuse or escalation.