# From Speech to Editable Concepts: Probing Emotion Recognition with Concept Bottleneck Models

arXiv: https://arxiv.org/abs/2609.39453 · HTML: https://mlsanity.com/paper/2609.39453/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 32 references that are in the map.
- First posted on arXiv 30 Sep 2026.

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

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting](https://mlsanity.com/paper/2305.04388/) (2023)
- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [Qwen2.5-Omni Technical Report](https://mlsanity.com/paper/2503.20215/) (2025)
- [MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversations](https://mlsanity.com/paper/1810.02508/) (2018)
- [Emotion Recognition from Speech Using Wav2vec 2.0 Embeddings](https://mlsanity.com/paper/2104.03502/) (2021)
- [Qwen3-ASR Technical Report](https://mlsanity.com/paper/2601.21337/) (2026)
- [A Closer Look at the Intervention Procedure of Concept Bottleneck Models](https://mlsanity.com/paper/2302.14260/) (2023)
- [Multimodal Speech Emotion Recognition Using Audio and Text](https://mlsanity.com/paper/1810.04635/) (2018)
- [EmoBox: Multilingual Multi-corpus Speech Emotion Recognition Toolkit and Benchmark](https://mlsanity.com/paper/2406.07162/) (2024)
- [BLSP-Emo: Towards Empathetic Large Speech-Language Models](https://mlsanity.com/paper/2406.03872/) (2024)
- [Beyond Silent Letters: Amplifying LLMs in Emotion Recognition with Vocal Nuances](https://mlsanity.com/paper/2407.21315/) (2024)
- [EmotionThinker: Prosody-Aware Reinforcement Learning for Explainable Speech Emotion Reasoning](https://mlsanity.com/paper/2601.15668/) (2026)
- [Using Auxiliary Tasks In Multimodal Fusion Of Wav2vec 2.0 And BERT For Multimodal Emotion Recognition](https://mlsanity.com/paper/2302.13661/) (2023)
- [VowelPrompt: Hearing Speech Emotions from Text via Vowel-level Prosodic Augmentation](https://mlsanity.com/paper/2602.06270/) (2026)
- [VoxEmo: Benchmarking Speech Emotion Recognition with Speech LLMs](https://mlsanity.com/paper/2603.08936/) (2026)

## Abstract (arXiv)

Speech emotion recognition (SER) is the task of assigning emotion labels to utterances. Early systems relied on acoustic features, whereas recent approaches combine multiple modalities, most commonly speech and text. Still, performance remains poor on many datasets. Large language models (LLMs) have therefore attracted interest for SER, as they can process diverse inputs jointly with instructions. However, direct audio input raises questions of explainability. To address similar questions in image classification, concept bottleneck models were introduced. This work adapts concept bottlenecks to SER to examine how individual predictions depend on transcripts, acoustic descriptions and speaker attributes. Experiments test three LLMs on CREMA-D, IEMOCAP and MELD, with concepts extracted by separate tools. On scripted corpora, LLMs are strongly biased towards the transcript in the zero-shot setting, which lowers Macro-F1 from 27.8 to 5.8 on CREMA-D. Fine-tuning removes this bias, and the transcript raises Macro-F1 from 41.8 to 45.1. Removing speech rate changes 48% of Neutral predictions to Disgust on CREMA-D; removing intensity level on MELD changes predictions despite little change in Macro-F1. These findings show that aggregate performance changes alone do not capture the effects of concept removal on individual predictions.