# Not in Sync: Unveiling Temporal Bias in Audio Chat Models

arXiv: https://arxiv.org/abs/2510.12185 · HTML: https://mlsanity.com/paper/2510.12185/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Audio-language model reasoning, one of 17 directions in LLM reasoning.
- Builds on 18 of its 20 references that are in the map.
- First posted on arXiv 14 Oct 2025.

## Placed in

- LLM reasoning / [Audio-language model reasoning](https://mlsanity.com/lr/audio-language-model-reasoning/)

## Builds on (most cited first, 18 of 18)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Lost in the Middle: How Language Models Use Long Contexts](https://mlsanity.com/paper/2307.03172/) (2023)
- [Survey of Hallucination in Natural Language Generation](https://mlsanity.com/paper/2202.03629/) (2022)
- [NExT-QA:Next Phase of Question-Answering to Explaining Temporal Actions](https://mlsanity.com/paper/2105.08276/) (2021)
- [Localizing Moments in Video with Natural Language](https://mlsanity.com/paper/1708.01641/) (2017)
- [AudioLM: a Language Modeling Approach to Audio Generation](https://mlsanity.com/paper/2209.03143/) (2022)
- [Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models](https://mlsanity.com/paper/2311.07919/) (2023)
- [TALL: Temporal Activity Localization via Language Query](https://mlsanity.com/paper/1705.02101/) (2017)
- [PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition](https://mlsanity.com/paper/1912.10211/) (2019)
- [SALMONN: Towards Generic Hearing Abilities for Large Language Models](https://mlsanity.com/paper/2310.13289/) (2023)
- [TVQA: Localized, Compositional Video Question Answering](https://mlsanity.com/paper/1809.01696/) (2018)
- [AudioGPT: Understanding and Generating Speech, Music, Sound, and Talking Head](https://mlsanity.com/paper/2304.12995/) (2023)
- [Sound Event Localization and Detection of Overlapping Sources Using Convolutional Recurrent Neural Networks](https://mlsanity.com/paper/1807.00129/) (2018)
- [Convolutional Recurrent Neural Networks for Polyphonic Sound Event Detection](https://mlsanity.com/paper/1702.06286/) (2017)
- [STARSS22: A dataset of spatial recordings of real scenes with spatiotemporal annotations of sound events](https://mlsanity.com/paper/2206.01948/) (2022)
- [Effective Pre-Training of Audio Transformers for Sound Event Detection](https://mlsanity.com/paper/2409.09546/) (2024)
- [Who is in the Spotlight: The Hidden Bias Undermining Multimodal Retrieval-Augmented Generation](https://mlsanity.com/paper/2506.11063/) (2025)
- [DCASE 2024 Task 4: Sound Event Detection with Heterogeneous Data and Missing Labels](https://mlsanity.com/paper/2406.08056/) (2024)

## Built on by (most cited first, 5 of 5)

- [Gated Differentiable Working Memory for Long-Context Language Modeling](https://mlsanity.com/paper/2601.12906/) (2026)
- [MusTBench: Benchmarking and Advancing Temporal Grounding in Music LLMs](https://mlsanity.com/paper/2605.29300/) (2026)
- [A Closer Look at Failure Modes in Temporal Understanding of Large Audio-Language Models](https://mlsanity.com/paper/2606.17417/) (2026)
- [Steering Where to Listen: Instruction-Based Activation Steering Redirects Temporal Attention in Large Audio-Language Models](https://mlsanity.com/paper/2606.11400/) (2026)
- [SPEARBench: A Benchmark for Naturalness Evaluation in Streaming Speech-to-Speech Language Models](https://mlsanity.com/paper/2607.05365/) (2026)

## Abstract (arXiv)

Large Audio Language Models (LALMs) are increasingly applied to audio understanding and multimodal reasoning, yet their ability to locate when events occur remains underexplored. We present the first systematic study of temporal bias in LALMs, revealing a key limitation in their timestamp prediction. For example, when asked "At which second does the lecturer introduce the key formula?", models often predict timestamps that are consistently earlier or later than the ground truth. Through controlled experiments on timestamped datasets, we find that temporal bias (i) is prevalent across datasets and models, (ii) increases with audio length - even accumulating to tens of seconds in extended recordings, and (iii) varies across event types and positions. We quantify this effect with the Temporal Bias Index (TBI), measuring systematic misalignment in predicted event timings, and complement it with a visualization framework. Our findings highlight a fundamental limitation in current LALMs and call for the development of temporally robust architectures.