# Measuring Audio's Impact on Correctness: Audio-Contribution-Aware Post-Training of Large Audio Language Models

arXiv: https://arxiv.org/abs/2509.21060 · HTML: https://mlsanity.com/paper/2509.21060/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 Sep 2025.

## Built on by (most cited first, 15 of 15)

- [All That Glitters Is Not Audio: Rethinking Text Priors and Audio Reliance in Audio-Language Evaluation](https://mlsanity.com/paper/2604.24401/) (2026)
- [Covo-Audio Technical Report](https://mlsanity.com/paper/2602.09823/) (2026)
- [Walking Through Uncertainty: An Empirical Study of Uncertainty Estimation for Audio-Aware Large Language Models](https://mlsanity.com/paper/2604.25591/) (2026)
- [DIFFA-2: A Practical Diffusion Large Language Model for General Audio Understanding](https://mlsanity.com/paper/2601.23161/) (2026)
- [AudioRouter: Data Efficient Audio Understanding via RL based Dual Reasoning](https://mlsanity.com/paper/2602.10439/) (2026)
- [A Closer Look at Failure Modes in Temporal Understanding of Large Audio-Language Models](https://mlsanity.com/paper/2606.17417/) (2026)
- [Audio-DeepThinker: Progressive Reasoning-Aware Reinforcement Learning for High-Quality Chain-of-Thought Emergence in Audio Language Models](https://mlsanity.com/paper/2604.18187/) (2026)
- [Is Text All You Need? Text as a Universal Information Bottleneck for Speech LLMs](https://mlsanity.com/paper/2606.09366/) (2026)
- [Tracing Audio Grounding and Answer Selection in Audio LLMs](https://mlsanity.com/paper/2609.04637/) (2026)
- [AudioDER: A Deduplication-Enhanced Reasoning Dataset for Post-Training Large Audio-Language Models](https://mlsanity.com/paper/2606.14591/) (2026)
- [Leading-Silence Augmentation and Multi-Stage Synthetic Supervision for the Second MLC-SLM Challenge](https://mlsanity.com/paper/2608.14150/) (2026)
- [SEAR: Segment-Evidence-Aware Routing for Weak-to-Strong Multilingual Speech MCQ](https://mlsanity.com/paper/2609.11355/) (2026)
- [SEABED: SouthEast Asian Benchmark for Evaluating Audio Reasoning](https://mlsanity.com/paper/2609.22586/) (2026)
- [Reward-Tilted On-Policy Distillation for Acoustic Grounding in Audio-Language Models](https://mlsanity.com/paper/2609.28778/) (2026)
- [Do Audio LLMs Listen Before They Act? Diagnosing Acoustic-Context Gating in Voice Agents](https://mlsanity.com/paper/2609.32536/) (2026)

## Abstract (arXiv)

Large Audio Language Models (LALMs) represent an important frontier in multimodal AI, addressing diverse audio tasks. Recently, post-training of LALMs has received increasing attention due to significant performance improvements over foundation models. While single-stage post-training such as reinforcement learning (RL) has demonstrated promising results, multi-stage approaches such as supervised fine-tuning (SFT) followed by RL remain suboptimal. The allocation of data across multiple training stages to maximize LALM capabilities has not been fully explored, and large-scale, high-quality datasets for such research are also lacking. To address these problems, we firstly present AudioMCQ, a comprehensive audio multiple-choice question dataset comprising 571k samples with two kinds of chain-of-thought annotations. Secondly, we investigate the prevalent zero audio-contribution phenomenon in LALMs, where models derive correct answers solely from textual information without processing audio content. We propose Audio-Contribution Filtering to partition data into weak and strong audio-contribution subsets. Based on these insights, we develop two effective post-training paradigms: Weak-to-Strong (SFT on weak audio-contribution data followed by RL on strong audio-contribution data) and Mixed-to-Strong (SFT on mixed audio-contribution data followed by RL on strong audio-contribution data). We achieve first place in the DCASE 2025 Audio-Question-Answering challenge by using AudioMCQ. Additionally, leveraging our dataset with different training strategies, we achieve 78.2% on MMAU-test-mini, 75.6% on MMAU, 67.1% on MMAR, and 70.7% on MMSU, establishing new state-of-the-art performance.