The Dominance of Text Space: Unveiling the Asymmetric Nature of Cross-Modal Alignment in Large Language Models (2026.acl-long)
Copied to clipboard
| Challenge: | Existing methods for cross-modal alignment assume a symmetric interaction between visual and textual modalities, implying that both spaces adapt to each other. |
| Approach: | They propose a method that regularizes the projector to maintain the geometric structure of the text embedding space via spectral filtering. |
| Outcome: | The proposed method preserves the LLM’s inherent linguistic capabilities and reduces object hallucination significantly better than standard fine-tuning methods. |
Similar Papers
Steering into New Embedding Spaces: Analyzing Cross-Lingual Alignment Induced by Model Interventions in Multilingual Language Models (2025.acl-long)
Copied to clipboard
Anirudh Sundar, Sinead Williamson, Katherine Metcalf, Barry-John Theobald, Skyler Seto, Masha Fedzechkina
| Challenge: | Large language models (LLMs) exhibit impressive performance on a variety of tasks from text summarization to zero-shot common-sense reasoning. |
| Approach: | They propose to manipulate the embedding space of mLLMs by manipulating its activations to steer generation into the desired direction. |
| Outcome: | The proposed model interventions improves alignment of cross-lingual representations in multilingual large language models with up to 2x improvements in top-1 accuracy on cross-linguistic retrieval tasks. |
Cross-Modal Projection in Multimodal LLMs Doesn’t Really Project Visual Attributes to Textual Space (2024.acl-short)
Copied to clipboard
| Challenge: | Existing multimodal large language models are limited to general-purpose multimodal tasks like question-answering on natural images. |
| Approach: | They propose to use cross-modal projection networks and a large language model to model domain-specific visual attributes of MLLMs. |
| Outcome: | The proposed models gain domain-specific visual capabilities when the projection is fine-tuned, but the updates do not extract relevant domain-specific visual attributes. |
Re-Align: Aligning Vision Language Models via Retrieval-Augmented Direct Preference Optimization (2025.emnlp-main)
Copied to clipboard
Shuo Xing, Peiran Li, Yuping Wang, Ruizheng Bai, Yueqi Wang, Chan-Wei Hu, Chengxuan Qian, Huaxiu Yao, Zhengzhong Tu
| Challenge: | emergence of large Vision Language Models (VLMs) has broadened the capabilities of single-modal Large Language Model (LLM) but VLMs are prone to significant hallucinations, especially in the form of cross-modal inconsistencies. |
| Approach: | They propose a new alignment framework that leverages image retrieval to integrate both textual and visual preference signals. |
| Outcome: | The proposed framework mitigates hallucinations more effectively than previous methods . it maintains robustness and scalability across a wide range of VLM sizes and architectures . |
Textual Steering Vectors Can Improve Visual Understanding in Multimodal Large Language Models (2026.acl-long)
Copied to clipboard
Woody Haosheng Gan, Deqing Fu, Julian Asilis, Ollie Liu, Vatsal Sharan, Robin Jia, Willie Neiswanger
| Challenge: | Steering methods have emerged as effective tools for guiding large language models’ behavior, yet multimodal large language model lacks comparable techniques due to architectural diversity and limited availability of multimodal steering vectors. |
| Approach: | They validate steering vectors derived solely from text-only LLM backbones and use a cross-modal transfer technique to reuse existing interpretability tools. |
| Outcome: | The proposed steering vectors can guide and enhance multimodal models using SPAR, Mean Shift, and Linear Probing. |
Multimodal Large Language Models for Text-rich Image Understanding: A Comprehensive Review (2025.findings-acl)
Copied to clipboard
Pei Fu, Tongkun Guan, Zining Wang, Zhentao Guo, Chen Duan, Hao Sun, Boming Chen, Qianyi Jiang, Jiayao Ma, Kai Zhou, Junfeng Luo
| Challenge: | Recent advances in vision-language models have unified perception and understanding tasks within Visual Question Answering paradigms. |
| Approach: | They propose to outline timeline, architecture, and pipeline of nearly all TIU MLLMs and review their performance on mainstream benchmarks. |
| Outcome: | The proposed models perform well on mainstream benchmarks and are compared with other models. |
Representational Isomorphism and Alignment of Multilingual Large Language Models (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing isomorphism of sentence representations can facilitate representational alignments in zero-shot and few-shot settings. |
| Approach: | They propose to apply a contrastive objective to LLMs with a small number of translation pairs to improve models' performance on Semantic Textual Similarity tasks. |
| Outcome: | The proposed representation-level approach significantly improves on Semantic Textual Similarity (STS) tasks across languages even without a monolingual objective. |
Concept Space Alignment in Multilingual LLMs (2024.emnlp-main)
Copied to clipboard
| Challenge: | Multilingual large language models generalize somewhat across languages, but it is unclear whether this is a result of improved, implicit alignment, or of something else, e.g., linguistic overlap or semi-parallel subsets of training data. |
| Approach: | They hypothesize that implicit alignment is the reason for generalization in multilingual large language models. |
| Outcome: | The proposed model generalizes well across languages, but lacks linearity. |
Generative Giants, Retrieval Weaklings: Why do Multimodal Large Language Models Fail at Multimodal Retrieval? (2026.findings-acl)
Copied to clipboard
| Challenge: | Rapid advances in multimodal large language models have revolutionized cross-modality understanding. |
| Approach: | They propose a method that uses whitening transformations to adjust MLLM representation spaces . they propose ML models that are dominated by textual semantics and visual semantics . |
| Outcome: | The proposed approach improves zero-shot multimodal retrieval performance without fine-tuning efforts. |
A Text is Worth Several Tokens: Text Embedding from LLMs Secretly Aligns Well with The Key Tokens (2025.acl-long)
Copied to clipboard
| Challenge: | et al., 2023) show that text embeddings from large language models can be aligned with key tokens in input text. |
| Approach: | They propose a sparse retrieval method based on aligned tokens for large language models . they show that this phenomenon is universal and is not affected by model architecture . |
| Outcome: | The proposed method can achieve 80% of the dense retrieval effect of the same model while reducing the computation significantly. |
VFA: Empowering Multilingual MLLMs via Vision-Free Adaptation (2026.acl-long)
Copied to clipboard
Yixia Li, Yaqing Shi, Zhiwen Ruan, Dongdong Zhang, Lingjie Jiang, Shaohan Huang, Yun Chen, Guanhua Chen, Furu Wei
| Challenge: | Multimodal large language models have advanced rapidly, yet most remain English-centric . scaling multilingual multimodal instruction tuning is limited by the scarcity and high cost of non-English image–text supervision. |
| Approach: | They propose a framework that decouples multilingual language enhancement from visual alignment by composing complementary task vectors over a shared LLM backbone. |
| Outcome: | The proposed framework achieves competitive performance with a fully multimodally trained model using less than 2% of the text data. |