MNAFT: modality neuron-aware fine-tuning of multimodal large language models for image translation

• URL: http://arxiv.org/abs/2604.16943v1
• Date: Sat, 18 Apr 2026 09:54:20 GMT
• Title: MNAFT: modality neuron-aware fine-tuning of multimodal large language models for image translation
• Abstract summary: Multimodal large language models (MLLMs) have shown impressive capabilities, yet they often struggle to effectively capture the fine-grained textual information within images.<n>Existing methods, primarily relying on instruction fine-tuning, risk parameter redundancy of pre-trained knowledge, hindering generalization performance.<n>We introduce modality neuron-aware fine-tuning (MNAFT), a novel approach that takes advantage of the specialized roles of individual neurons within MLLMs for enhanced image translation.
• Score: 19.679466742536892
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Multimodal large language models (MLLMs) have shown impressive capabilities, yet they often struggle to effectively capture the fine-grained textual information within images crucial for accurate image translation. This often leads to a modality gap between visual text inputs and textual inputs/outputs for image translation. Existing methods, primarily relying on instruction fine-tuning, risk parameter redundancy of pre-trained knowledge, hindering generalization performance. To address this, we introduce modality neuron-aware fine-tuning (MNAFT), a novel approach that takes advantage of the specialized roles of individual neurons within MLLMs for enhanced image translation. MNAFT identifies language-agnostic and language-specific neurons in both vision and language modules through an instruction-driven activation analysis, evaluating their importance in various translation tasks. We then perform selective fine-tuning, updating only the parameters of language-specific and language-agnostic neurons within the selected layers relevant to the target task, while preserving the knowledge encoded in other neurons and layers. Our extensive experiments on multiple benchmarks demonstrate that MNAFT significantly outperforms state-of-the-art image translation methods, including cascaded models, standard full fine-tuning, and parameter-efficient tuning techniques. Furthermore, we provide comprehensive analysis, including visualizations of neuron activations and clustering patterns, to offer insights into the roles of different neuron groups in mediating cross-modal understanding and facilitating accurate language-specific translation.

Related papers

• Beyond Language Modeling: An Exploration of Multimodal Pretraining [125.34714978184638]
  We provide empirical clarity through controlled, from-scratch pretraining experiments.<n>We adopt the Transfusion framework, using next-token prediction for language and diffusion for vision.<n>We demonstrate that the MoE architecture harmonizes this scaling asymmetry by providing the high model capacity required by language.
  arXiv  Detail & Related papers  (2026-03-03T18:58:00Z)
• Language-Specific Layer Matters: Efficient Multilingual Enhancement for Large Vision-Language Models [60.39744129890118]
  Large vision-language models (LVLMs) have demonstrated exceptional capabilities in understanding visual information with human languages.<n>In this work, we identify a salient correlation between the multilingual understanding ability of LVLMs and language-specific neuron activations in shallow layers.<n>We introduce PLAST, a training recipe that achieves efficient multilingual enhancement for LVLMs by Precise LAnguage-Specific layers fine-Tuning.
  arXiv  Detail & Related papers  (2025-08-25T18:15:25Z)
• LLaVA-NeuMT: Selective Layer-Neuron Modulation for Efficient Multilingual Multimodal Translation [12.51212639515934]
  LLaVA-NeuMT is a novel framework that explicitly models language-specific and language-agnostic representations to mitigate multilingual interference.<n>Our approach consists of a layer selection mechanism that identifies the most informative layers for different language pairs.<n>We conduct extensive experiments on the M3-Multi30K and M3-AmbigCaps datasets, demonstrating that LLaVA-NeuMT, while fine-tuning only 40% of the model parameters, surpasses full fine-tuning approaches.
  arXiv  Detail & Related papers  (2025-07-25T04:23:24Z)
• Locate-then-Merge: Neuron-Level Parameter Fusion for Mitigating Catastrophic Forgetting in Multimodal LLMs [19.472889262384818]
  We propose a training-free parameter fusion framework that locates important parameters and then selectively merges them.<n>We introduce Neuron-Fusion, a neuron-level strategy that preserves the influence of neurons with large parameter shifts.<n>Experiments on 13 benchmarks across both language and visual tasks show that Neuron-Fusion consistently outperforms existing model merging methods.
  arXiv  Detail & Related papers  (2025-05-22T14:04:43Z)
• LANDeRMT: Detecting and Routing Language-Aware Neurons for Selectively Finetuning LLMs to Machine Translation [43.26446958873554]
  Large language models (LLMs) have shown promising results in multilingual translation even with limited bilingual supervision. Recent advancements in large language models (LLMs) have shown promising results in multilingual translation even with limited bilingual supervision. LandeRMT is a framework that selectively finetunes LLMs to textbfMachine textbfTranslation with diverse translation training data.
  arXiv  Detail & Related papers  (2024-09-29T02:39:42Z)
• Modelling Multimodal Integration in Human Concept Processing with Vision-Language Models [7.511284868070148]
  We investigate whether integration of visuo-linguistic information leads to representations that are more aligned with human brain activity.<n>Our findings indicate an advantage of multimodal models in predicting human brain activations.
  arXiv  Detail & Related papers  (2024-07-25T10:08:37Z)
• Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization [52.935150075484074]
  We introduce a well-designed visual tokenizer to translate the non-linguistic image into a sequence of discrete tokens like a foreign language. The resulting visual tokens encompass high-level semantics worthy of a word and also support dynamic sequence length varying from the image. This unification empowers LaVIT to serve as an impressive generalist interface to understand and generate multi-modal content simultaneously.
  arXiv  Detail & Related papers  (2023-09-09T03:01:38Z)
• Multimodal Neurons in Pretrained Text-Only Transformers [52.20828443544296]
  We identify "multimodal neurons" that convert visual representations into corresponding text. We show that multimodal neurons operate on specific visual concepts across inputs, and have a systematic causal effect on image captioning.
  arXiv  Detail & Related papers  (2023-08-03T05:27:12Z)
• SNeL: A Structured Neuro-Symbolic Language for Entity-Based Multimodal Scene Understanding [0.0]
  We introduce SNeL (Structured Neuro-symbolic Language), a versatile query language designed to facilitate nuanced interactions with neural networks processing multimodal data. SNeL's expressive interface enables the construction of intricate queries, supporting logical and arithmetic operators, comparators, nesting, and more. Our evaluations demonstrate SNeL's potential to reshape the way we interact with complex neural networks.
  arXiv  Detail & Related papers  (2023-06-09T17:01:51Z)
• Universal Multimodal Representation for Language Understanding [110.98786673598015]
  This work presents new methods to employ visual information as assistant signals to general NLP tasks. For each sentence, we first retrieve a flexible number of images either from a light topic-image lookup table extracted over the existing sentence-image pairs. Then, the text and images are encoded by a Transformer encoder and convolutional neural network, respectively.
  arXiv  Detail & Related papers  (2023-01-09T13:54:11Z)
• A Novel Graph-based Multi-modal Fusion Encoder for Neural Machine Translation [131.33610549540043]
  We propose a novel graph-based multi-modal fusion encoder for NMT. We first represent the input sentence and image using a unified multi-modal graph. We then stack multiple graph-based multi-modal fusion layers that iteratively perform semantic interactions to learn node representations.
  arXiv  Detail & Related papers  (2020-07-17T04:06:09Z)
• Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation [81.7786241489002]
  Massively multilingual models for neural machine translation (NMT) are theoretically attractive, but often underperform bilingual models and deliver poor zero-shot translations. We argue that multilingual NMT requires stronger modeling capacity to support language pairs with varying typological characteristics. We propose random online backtranslation to enforce the translation of unseen training language pairs.
  arXiv  Detail & Related papers  (2020-04-24T17:21:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.