mlsanityData to 14 Oct 2026

Investigating Faithfulness in Large Audio Language Models

arXiv 2509.22363 ↗First posted 26 Sep 2025cs.LG eess.AScited by 1

On the map: LLM reasoning → Audio-language model reasoning

Abstract (arXiv)

Large Audio Language Models (LALMs) integrate audio encoders with pretrained Large Language Models to perform complex multimodal reasoning tasks. While these models can generate Chain-of-Thought (CoT) explanations, the faithfulness of these reasoning chains remains unclear. In this work, we propose a systematic framework to evaluate CoT faithfulness in LALMs with respect to both the input audio and the final model prediction. We define three criteria for audio faithfulness: hallucination-free, holistic, and attentive listening. We also introduce a benchmark based on both audio and CoT interventions to assess faithfulnessfootnoteThe benchmarking interface and evaluation results are available at https://poonehmousavi.github.io/faithfulness/. Experiments on Audio Flamingo 3 and Qwen2.5-Omni suggest a potential multimodal disconnect: reasoning often aligns with the final prediction but is not always strongly grounded in the audio and can be vulnerable to hallucinations or adversarial perturbations.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Audio-language model reasoning, one of 17 directions in LLM reasoning. Builds on 37 of its 42 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

37 of 42 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025Audio Flamingo Sound-CoT Technical Report: Improving Chain-of-Thought Reasoning in Sound Understanding15 Aug
  2. 2025Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models10 Jul
  3. 2025SAKURA: On the Multi-hop Reasoning of Large Audio-Language Models Based on Speech and Audio Information19 May
  4. 2025MMAR: A Challenging Benchmark for Deep Reasoning in Speech, Audio, Music, and Their Mix19 May
  5. 2025Qwen2.5-Omni Technical Report26 Mar
  6. 2025Audio-Reasoner: Improving Reasoning Capability in Large Audio Language Models4 Mar
  7. 2025Audio-CoT: Exploring Chain-of-Thought Reasoning in Large Audio Language Model13 Jan
  8. 2024MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark24 Oct
  9. 2024What Are They Doing? Joint Audio-Speech Co-Reasoning22 Sep

Other references · 28

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  4. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  5. 2023Qwen Technical Reportcited by 3,344
  6. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,211
  7. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  8. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Modelscited by 1,407
  9. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Promptingcited by 1,122
  10. 2023Measuring Faithfulness in Chain-of-Thought Reasoningcited by 703
  11. 2024Qwen2-Audio Technical Reportcited by 560
  12. 2023SALMONN: Towards Generic Hearing Abilities for Large Language Modelscited by 506
16 more
  1. 2025Qwen3-Omni Technical Reportcited by 497
  2. 2023OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocationcited by 478
  3. 2024CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Modelscited by 288
  4. 2025Chain-of-Thought Reasoning In The Wild Is Not Always Faithfulcited by 170
  5. 2017What do we need to build explainable AI systems for the medical domain?cited by 129
  6. 2022Explanations from Large Language Models Make Small Reasoners Bettercited by 124
  7. 2025On The Landscape of Spoken Language Models: A Comprehensive Surveycited by 118
  8. 2022Rationale-Augmented Ensembles in Language Modelscited by 111
  9. 2025Step-Audio 2 Technical Reportcited by 105
  10. 2024Are self-explanations from Large Language Models faithful?cited by 101
  11. 2023Can Large Language Models Explain Themselves? A Study of LLM-Generated Self-Explanationscited by 90
  12. 2023Do Models Explain Themselves? Counterfactual Simulatability of Natural Language Explanationscited by 85
  13. 2025Walk the Talk? Measuring the Faithfulness of Large Language Model Explanationscited by 45
  14. 2025Mirage in the Eyes: Hallucination Attack on Multi-modal Large Language Models with Only Attention Sinkcited by 15
  15. 2025Evaluating Robustness of Large Audio Language Models to Audio Injection: An Empirical Studycited by 12
  16. 2025LiSTEN: Learning Soft Token Embeddings for Neural Audio LLMscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Acoustic Cue Alignment in Audio Language Models for Speech Emotion Recognitioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Can Large Audio-Language Models Truly Hear? Tackling Hallucinations with Multi-Task Assessment and Stepwise Audio Reasoning
  2. 2025Thinking with Sound: Audio Chain-of-Thought Enables Multimodal Reasoning in Large Audio-Language Models
  3. 2025AHA: Aligning Large Audio-Language Models for Reasoning Hallucinations via Counterfactual Hard Negatives
  4. 2026Audio-Cogito: Towards Deep Audio Reasoning in Large Audio Language Models
  5. 2025SightSound-R1: Cross-Modal Reasoning Distillation from Vision to Audio Language Models
  6. 2026A Benchmark for Audio Reasoning Capabilities of Multimodal Large Language Models
  7. 2026A Closer Look at Failure Modes in Temporal Understanding of Large Audio-Language Models
  8. 2026AudioRouter: Data Efficient Audio Understanding via RL based Dual Reasoning