# Delta-AttnMask: Attention-Guided Masked Hidden States for Efficient Data Selection and Augmentation

arXiv: https://arxiv.org/abs/2508.09199 · HTML: https://mlsanity.com/paper/2508.09199/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in MLLM training and visual instruction tuning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 21 of its 31 references that are in the map.
- First posted on arXiv 8 Aug 2025.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM training and visual instruction tuning](https://mlsanity.com/mllm/mllm-training-and-visual-instruction-tuning/)

## Builds on (most cited first, 21 of 21)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering](https://mlsanity.com/paper/2209.09513/) (2022)
- [LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models](https://mlsanity.com/paper/2403.13372/) (2024)
- [SEED-Bench: Benchmarking Multimodal LLMs with Generative Comprehension](https://mlsanity.com/paper/2307.16125/) (2023)
- [LESS: Selecting Influential Data for Targeted Instruction Tuning](https://mlsanity.com/paper/2402.04333/) (2024)
- [From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuning](https://mlsanity.com/paper/2308.12032/) (2023)
- [A Survey on Vision-Language-Action Models for Embodied AI](https://mlsanity.com/paper/2405.14093/) (2024)
- [Prioritized Training on Points that are Learnable, Worth Learning, and Not Yet Learnt](https://mlsanity.com/paper/2206.07137/) (2022)
- [Online Batch Selection for Faster Training of Neural Networks](https://mlsanity.com/paper/1511.06343/) (2015)
- [Accelerating Deep Learning by Focusing on the Biggest Losers](https://mlsanity.com/paper/1910.00762/) (2019)
- [Your Vision-Language Model Itself Is a Strong Filter: Towards High-Quality Instruction Tuning with Data Selection](https://mlsanity.com/paper/2402.12501/) (2024)
- [ICONS: Influence Consensus for Vision-Language Data Selection](https://mlsanity.com/paper/2501.00654/) (2024)
- [A CLIP-Powered Framework for Robust and Generalizable Data Selection](https://mlsanity.com/paper/2410.11215/) (2024)
- [Filter Images First, Generate Instructions Later: Pre-Instruction Data Selection for Visual Instruction Tuning](https://mlsanity.com/paper/2503.07591/) (2025)
- [Mastering Collaborative Multi-modal Data Selection: A Focus on Informativeness, Uniqueness, and Representativeness](https://mlsanity.com/paper/2412.06293/) (2024)
- [Curriculum Learning with Quality-Driven Data Selection](https://mlsanity.com/paper/2407.00102/) (2024)
- [Sensitivity Analysis of Deep Neural Networks](https://mlsanity.com/paper/1901.07152/) (2019)
- [Better Reasoning with Less Data: Enhancing VLMs Through Unified Modality Scoring](https://mlsanity.com/paper/2506.08429/) (2025)
- [TinyAlign: Boosting Lightweight Vision-Language Models by Mitigating Modal Alignment Bottlenecks](https://mlsanity.com/paper/2505.12884/) (2025)

## Abstract (arXiv)

Visual Instruction Finetuning (VIF) is pivotal for post-training Vision-Language Models (VLMs). Unlike unimodal instruction finetuning in plain-text large language models, which mainly requires instruction datasets to enable model instruction-following ability, VIF also requires multimodal data to enable joint visual and textual understanding; therefore, it typically requires more data. Consequently, VIF imposes stricter data selection challenges: the method must scale efficiently to handle larger data demands while ensuring the quality of both visual and textual content, as well as their alignment. Despite its critical impact on performance, data selection for VIF remains an understudied area. In this paper, we propose Delta-AttnMask. This data-efficient framework quantifies sample quality through attention-guided masking of the model's hidden states, jointly evaluating image-text pairs without requiring domain labels, auxiliary models, or extra training. By computing loss differences (Delta) between the original states and states masked using high-attention regions, Delta-AttnMask intrinsically assesses sample quality. Experiments across multiple VLMs and datasets show that Delta-AttnMask achieves state-of-the-art performance with just 20% of data, accelerating training by 5x while surpassing full-dataset baselines by +10.1% in overall accuracy. Its model-agnostic and data-agnostic design ensures broad applicability across modalities and architectures.