# CALM: Class-Conditional Sparse Attention Vectors for Large Audio-Language Models

arXiv: https://arxiv.org/abs/2602.07077 · HTML: https://mlsanity.com/paper/2602.07077/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 34 of its 49 references that are in the map.
- First posted on arXiv 6 Feb 2026.

## Builds on (most cited first, 30 of 34)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification](https://mlsanity.com/paper/1709.00029/) (2017)
- [What Does BERT Look At? An Analysis of BERT's Attention](https://mlsanity.com/paper/1906.04341/) (2019)
- [Probing Classifiers: Promises, Shortcomings, and Advances](https://mlsanity.com/paper/2102.12452/) (2021)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [VGGSound: A Large-scale Audio-Visual Dataset](https://mlsanity.com/paper/2004.14368/) (2020)
- [Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs](https://mlsanity.com/paper/2503.01743/) (2025)
- [AudioCLIP: Extending CLIP to Image, Text and Audio](https://mlsanity.com/paper/2106.13043/) (2021)
- [In-Context Learning Creates Task Vectors](https://mlsanity.com/paper/2310.15916/) (2023)
- [AudioGPT: Understanding and Generating Speech, Music, Sound, and Talking Head](https://mlsanity.com/paper/2304.12995/) (2023)
- [Macaw-LLM: Multi-Modal Language Modeling with Image, Audio, Video, and Text Integration](https://mlsanity.com/paper/2306.09093/) (2023)
- [The Butterfly Effect of Altering Prompts: How Small Changes and Jailbreaks Affect Large Language Model Performance](https://mlsanity.com/paper/2401.03729/) (2024)
- [Fine-Tuned 'Small' LLMs (Still) Significantly Outperform Zero-Shot Generative AI Models in Text Classification](https://mlsanity.com/paper/2406.08660/) (2024)
- [Multimodal Task Vectors Enable Many-Shot Multimodal In-Context Learning](https://mlsanity.com/paper/2406.15334/) (2024)
- [Differentiation and Specialization of Attention Heads via the Refined Local Learning Coefficient](https://mlsanity.com/paper/2410.02984/) (2024)
- [WAVPROMPT: Towards Few-Shot Spoken Language Understanding with Frozen Language Models](https://mlsanity.com/paper/2203.15863/) (2022)
- [UniAudio 1.5: Large Language Model-driven Audio Codec is A Few-shot Audio Task Learner](https://mlsanity.com/paper/2406.10056/) (2024)
- [Head Pursuit: Probing Attention Specialization in Multimodal Transformers](https://mlsanity.com/paper/2510.21518/) (2025)
- [Zero-Shot Audio Classification Based on Class Label Embeddings](https://mlsanity.com/paper/1905.01926/) (2019)
- [Gen-Z: Generative Zero-Shot Text Classification with Contextualized Label Descriptions](https://mlsanity.com/paper/2311.07115/) (2023)
- [AQA-TTRL: Self-Adaptation in Audio Question Answering with Test-Time Reinforcement Learning](https://mlsanity.com/paper/2510.05478/) (2025)
- [Domain Adaptation for Contrastive Audio-Language Models](https://mlsanity.com/paper/2402.09585/) (2024)
- [PALM: Few-Shot Prompt Learning for Audio Language Models](https://mlsanity.com/paper/2409.19806/) (2024)
- [Exploring Fine-Tuning of Large Audio Language Models for Spoken Language Understanding under Limited Speech Data](https://mlsanity.com/paper/2509.15389/) (2025)
- [Enhancing Temporal Understanding in Audio Question Answering for Large Audio Language Models](https://mlsanity.com/paper/2409.06223/) (2024)
- [Cognitive Mirrors: Exploring the Diverse Functional Roles of Attention Heads in LLM Reasoning](https://mlsanity.com/paper/2512.10978/) (2025)
- [LEAN: Light and Efficient Audio Classification Network](https://mlsanity.com/paper/2305.12712/) (2023)

## Abstract (arXiv)

Large audio-language models (LALMs) exhibit strong zero-shot capabilities in multiple downstream tasks, such as audio question answering (AQA) and abstract reasoning; however, these models still lag behind specialized models for certain discriminative tasks (e.g., audio classification). Recent studies show that sparse subsets of attention heads within an LALM can serve as strong discriminative feature extractors for downstream tasks such as classification via simple voting schemes. However, these methods assign uniform weights to all selected heads, implicitly assuming that each head contributes equally across all semantic categories. In this work, we propose Class-Conditional Sparse Attention Vectors for Large Audio-Language Models, a few-shot classification method that learns class-dependent importance weights over attention heads. This formulation allows individual heads to specialize in distinct semantic categories and to contribute to ensemble predictions proportionally to their estimated reliability. Experiments on multiple few-shot audio and audiovisual classification benchmarks and tasks demonstrate that our method consistently outperforms state-of-the-art uniform voting-based approaches by up to 14.52%, 1.53%, 8.35% absolute gains for audio classification, audio-visual classification, and spoofing detection respectively.