arXiv Computation and Language October 26, 2026 By Jia Wang, Yuqing Li, Jiangnan Li, Zheng Lin, Mo Yu

A Broader Look into MLLM's Retrieval Heads

Read the original on arXiv Computation and Language →

The paper investigates how retrieval heads—sparse attention heads that transfer information from context to output—behave when multimodal large language models (MLLMs) are extended to include audio-visual inputs. Across 18 language, vision‑language, and omni‑modal checkpoints, the authors find that retrieval heads remain sparse and largely shared across modalities, with less than 2 % of heads being common to all modalities; ablating these shared heads degrades tasks such as document understanding, spatial perception, and visual mathematics. By training retrieval heads with a ranking objective, they demonstrate a 6.03‑point improvement in Recall@1 on the MMDocIR benchmark for Qwen3‑VL‑8B and improved document retrieval across three model families.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

arXiv AI

Jun 6

Mechanistic Insights into Functional Sparsity in Multimodal LLMs via CoRe Heads

arXiv:2606. 05843v1 Announce Type: cross Abstract: While Multimodal Large Language Models (MLLMs) demonstrate remarkable proficiency on complex vision-language tasks, the mechanisms by which they extract query-relevant visual features from complex, noisy contexts remain opaque.

By Ruoxi Sun, Quantong Qiu, Juntao Li, Zecheng Tang, Yihang Lou, Min Zhang

llms efficiency multimodal safety

arXiv Computer Vision

Sep 1

Can Retrieval Heads See Images? Multimodal Retrieval Heads in Long-Context Vision-Language Models

arXiv:2605.27243v3 Announce Type: replace Abstract: Large vision-language models increasingly rely on long-context modeling to reason over documents, hour-level videos, and long-horizon agent traject...

By Aaron Branson Cigres Li, Zhaowei Wang, Yu Zhao, Yiming Du, Haobo Li, Xiyu Ren, Ginny Wong, Simon See, Lishu Luo, Haodong Duan, Pasquale Minervini, Yangqiu Song

llms agents multimodal

arXiv Machine Learning

Jul 14

Empowering Long-form Omni-modal Understanding with Robust Audio Perception

arXiv:2607. 10299v1 Announce Type: new Abstract: Recent advances in large-scale multimodal models have drivenremarkable progress in vision-language tasks; however, comprehensiveomni-modal understanding remains under-explored, largely due to thescarcity of datasets with rich, explicitly aligned auditory cues.

By Kaiying Yan, Luoyi Sun, Xiao Zhou, Weidi Xie

fine-tuning multimodal benchmarks

arXiv Computer Vision

3d ago

The Answer Knows Where to Look: Zooming before MLLMs Answer under Visual-Token Budgets

arXiv:2610.32664v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) read images through visual tokens, and under the small per-image budgets that efficient deployment requires, t...

By Zhensheng Hu, Guanglong Xu, Guanqiang Hu

llms multimodal benchmarks

arXiv Machine Learning

Sep 30

Seeing What Should Be Heard: Diagnosing and Repairing Cross-Modal Shortcuts in Omni-Modal LLMs

arXiv:2609.36798v1 Announce Type: cross Abstract: Omni-modal large language models (LLMs) are expected to answer a question using the modality it explicitly refers to. However, existing training para...

By Yueran Ma, Ronghao Lin

llms nlp reinforcement-learning fine-tuning multimodal benchmarks

arXiv AI

Aug 11

Omni2LoRA: Coherence-Preserving Parametric Memory for Efficient Omni Language Models

arXiv:2608. 09227v1 Announce Type: new Abstract: Omnimodal language models (OLMs) enable unified audio-visual understanding, but processing long joint token sequences makes inference computationally prohibitive.

By Puneet Mathur, Manan Suri, Dinesh Manocha

llms nlp fine-tuning efficiency multimodal benchmarks