Salvato in:
Dettagli Bibliografici
Autori principali: Lee, Taehan, Jung, Jaehan, Lee, Hyukjun
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2603.03855
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910040388009984
author Lee, Taehan
Jung, Jaehan
Lee, Hyukjun
author_facet Lee, Taehan
Jung, Jaehan
Lee, Hyukjun
contents Audio LLMs have shown a strong ability to understand audio samples, yet their reliability in complex acoustic scenes remains under-explored. Unlike prior work limited to small scale or less controlled query construction, we present a large-scale evaluation of event grounding and false alarms as auditory scene complexity increases. Using 71K AudioCapsV2 clips, we extract normalized (source, attribute) events and build two query types: present-event queries for ground-truth detection and absent-event queries to probe hallucinations, using similarity-filtered negative sampling in an audio-aligned text embedding space. We evaluate four SOTA Audio LLMs with 12 prompt variants over 500K yes/no queries per model. Across models, increasing event count consistently lowers true-positive rate and raises false-positive rate, while prompts induce a strong trade-off between the two. Our confidence analysis shows that models become more uncertain on multi-event audio, revealing room for improvement.
format Preprint
id arxiv_https___arxiv_org_abs_2603_03855
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Sensitivity Analysis of Multi-Event Audio Grounding in Audio LLMs
Lee, Taehan
Jung, Jaehan
Lee, Hyukjun
Sound
Audio LLMs have shown a strong ability to understand audio samples, yet their reliability in complex acoustic scenes remains under-explored. Unlike prior work limited to small scale or less controlled query construction, we present a large-scale evaluation of event grounding and false alarms as auditory scene complexity increases. Using 71K AudioCapsV2 clips, we extract normalized (source, attribute) events and build two query types: present-event queries for ground-truth detection and absent-event queries to probe hallucinations, using similarity-filtered negative sampling in an audio-aligned text embedding space. We evaluate four SOTA Audio LLMs with 12 prompt variants over 500K yes/no queries per model. Across models, increasing event count consistently lowers true-positive rate and raises false-positive rate, while prompts induce a strong trade-off between the two. Our confidence analysis shows that models become more uncertain on multi-event audio, revealing room for improvement.
title A Sensitivity Analysis of Multi-Event Audio Grounding in Audio LLMs
topic Sound
url https://arxiv.org/abs/2603.03855