# Steer-MoE: Efficient Audio-Language Alignment with a Mixture-of-Experts Steering Module

arXiv: https://arxiv.org/abs/2510.13558 · HTML: https://mlsanity.com/paper/2510.13558/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal MoE models, one of 12 directions in Mixture-of-experts.
- Builds on 9 of its 14 references that are in the map.
- First posted on arXiv 15 Oct 2025.

## Placed in

- Mixture-of-experts / [Multimodal MoE models](https://mlsanity.com/moe/multimodal-moe-models/)

## Builds on (most cited first, 9 of 9)

- [Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation](https://mlsanity.com/paper/2211.06687/) (2022)
- [Qwen2-Audio Technical Report](https://mlsanity.com/paper/2407.10759/) (2024)
- [AISHELL-2: Transforming Mandarin ASR Research Into Industrial Scale](https://mlsanity.com/paper/1808.10583/) (2018)
- [WavLLM: Towards Robust and Adaptive Speech Large Language Model](https://mlsanity.com/paper/2404.00656/) (2024)
- [Clotho-AQA: A Crowdsourced Dataset for Audio Question Answering](https://mlsanity.com/paper/2204.09634/) (2022)
- [FireRedASR: Open-Source Industrial-Grade Mandarin Speech Recognition Models from Encoder-Decoder to LLM Integration](https://mlsanity.com/paper/2501.14350/) (2025)
- [Paralinguistics-Aware Speech-Empowered Large Language Models for Natural Conversation](https://mlsanity.com/paper/2402.05706/) (2024)
- [Cross-model Transferability among Large Language Models on the Platonic Representations of Concepts](https://mlsanity.com/paper/2501.02009/) (2025)
- [Teaching Physical Awareness to LLMs through Sounds](https://mlsanity.com/paper/2506.08524/) (2025)

## Built on by (most cited first, 3 of 3)

- [SALSA: Speech Aware LLM Adaptation via Learned Steering Activation Vectors](https://mlsanity.com/paper/2606.00460/) (2026)
- [AsymSpec: Context-Asymmetric Speculative Decoding for Agentic LLMs](https://mlsanity.com/paper/2608.26004/) (2026)
- [Repairing Speech Access After Text Updates with Frozen Affine Fields](https://mlsanity.com/paper/2610.20253/) (2026)

## Abstract (arXiv)

Aligning pretrained audio encoders and Large Language Models (LLMs) offers a promising, parameter-efficient path to building powerful multimodal agents. However, existing methods often require costly full-model finetuning or rely on static adapters that may lack expressive power. Drawing inspiration from the Platonic Representation Hypothesis, we introduce SteerMoE, a novel and modular framework for audio-language alignment. SteerMoE freezes both the audio encoder and the LLM decoder, training only a lightweight steering module integrated within the encoder's layers. This module uses a Mixture-of-Experts (MoE) router to dynamically select and apply learned steering vectors, progressively transforming continuous audio representations into a space comprehensible to the LLM. By operating entirely in the continuous embedding space, our approach requires no modifications to the LLM's vocabulary and preserves its advanced reasoning and agentic capabilities. We demonstrate through experiments on ASR, audio understanding, and a qualitative function-calling task that SteerMoE achieves strong performance while remaining highly modular and computationally efficient, offering a robust new paradigm for developing sophisticated audio-language systems.