Enhancing Multilingual Reasoning via Steerable Model Merging (2026.findings-acl)
Copied to clipboard
Zhuoran Li, Rui Xu, Jian Yang, Junnan Liu, Zhijun Chen, Qianren Mao, Hongcheng Guo, Jiaheng Liu, Likang Xiao, Ming LI, Xiaojie Wang
| Challenge: | Model merging is an effective technique for composing the capabilities of a multilingual model and a reasoning model. |
| Approach: | They propose a model merging framework that modulates the contribution of each source model. |
| Outcome: | Experiments show that the proposed model merging framework outperforms strong baselines on multilingual reasoning benchmarks across 21 different languages. |
Similar Papers
Merge to Learn: Efficiently Adding Skills to Language Models with Model Merging (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Adapting general-purpose language models to new skills is currently expensive . Adaptation to new skill sets requires repeated training or models forget older skills . |
| Approach: | They propose a parallel-train-then-merge procedure that adds new skills to preexisting models in isolation and later merges with the general model. |
| Outcome: | The proposed method is cheaper than retraining models on updated datasets . it improves model compliance with safe prompts while preserving model's ability to refuse dangerous or harmful prompts. |
MergeME: Model Merging Techniques for Homogeneous and Heterogeneous MoEs (2025.naacl-long)
Copied to clipboard
Yuhang Zhou, Giannis Karamanolakis, Victor Soto, Anna Rumshisky, Mayank Kulkarni, Furong Huang, Wei Ai, Jianhua Lu
| Challenge: | State-of-the-art methods for merging expert models with different architectures do not address parameter interference and require extensive fine-tuning to restore performance. |
| Approach: | They propose a method for merging experts with different architectures into a unified Mixture-of-Experts model with a goal of enhancing performance in each domain while retaining effectiveness on general tasks. |
| Outcome: | Experiments across multiple domains show that the proposed methods reduce fine-tuning costs and improve performance over state-of-the-art methods. |
Sens-Merging: Sensitivity-Guided Parameter Balancing for Merging Large Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing task vector-based model merging methods apply uniform coefficients across all parameters, overlooking varying parameter importance both within and across tasks. |
| Approach: | They propose a sensitivity-guided coefficient adjustment method that optimizes existing model merging techniques by operating at both task-specific and cross-task levels. |
| Outcome: | The proposed method outperforms existing model merging techniques on mistral 7B and LLaMA2 7B/13B models and enables them to outperformed specialized models. |
Split-Merge: Scalable and Memory-Efficient Merging of Expert LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | a zero-shot merging framework for large language models consolidates specialized domain experts into a single model without any further training. |
| Approach: | They propose a zero-shot merging framework that consolidates specialized domain experts into a single model without further training. |
| Outcome: | Experiments on code generation, mathematical reasoning, medical question answering, and instruction-following benchmarks confirm the versatility and effectiveness of the proposed framework. |
Training-free LLM Merging for Multi-task Learning (2025.acl-long)
Copied to clipboard
Zichuan Fu, Xian Wu, Yejing Wang, Wanyu Wang, Shanshan Ye, Hongzhi Yin, Yi Chang, Yefeng Zheng, Xiangyu Zhao
| Challenge: | Large Language Models (LLMs) have demonstrated exceptional capabilities across diverse natural language processing tasks. |
| Approach: | They propose a training-free method for unifying different specialized LLMs into a single model using model-wise and layer-wise pruning and scaling. |
| Outcome: | The proposed method outperforms existing merging techniques and surpasses models fine-tuned on combined datasets in most scenarios. |
Model Merging and Safety Alignment: One Bad Model Spoils the Bunch (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Existing methods for merging large language models often overlook safety alignment during merging, leading to misaligned models. |
| Approach: | They propose to combine safety and domain-specific data to optimize model merging techniques . they propose to use this data to maximize model alignment . |
| Outcome: | The proposed method allows for models that excel in both domain expertise and alignment. |
LayAlign: Enhancing Multilingual Reasoning in Large Language Models via Layer-Wise Adaptive Fusion and Alignment Strategy (2025.findings-naacl)
Copied to clipboard
| Challenge: | Large language models (LLMs) are pretrained on multilingual corpora but exhibit suboptimal performance on low-resource languages. |
| Approach: | They propose a framework that integrates representations from all encoder layers and an adaptive fusion-enhanced attention mechanism to enable layer-wise interaction between the LLM and the multilingual encoder. |
| Outcome: | Experiments on multilingual reasoning tasks show that the proposed framework outperforms baselines. |
LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging (2025.findings-emnlp)
Copied to clipboard
| Challenge: | a framework for model merging is proposed without additional training . task vectors from fine-tuned models exhibit a limited number of dominant singular values . |
| Approach: | They propose a framework for model merging based on low-rank estimation of task vectors without access to the base model. |
| Outcome: | The proposed framework improves models without additional training without additional inputs. |
UQ-Merge: Uncertainty Guided Multimodal Large Language Model Merging (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing models merging methods often lead to suboptimal performance due to harmful models . et al., 2018; 59: 59-64. |
| Approach: | They propose an uncertainty-guided MLLM merging algorithm that integrates models into a single MLML. |
| Outcome: | The proposed algorithm improves on held-in and held-out vision-language benchmarks. |
Multilingual Reasoning via Self-training (2025.naacl-long)
Copied to clipboard
| Challenge: | Recent studies have introduced eclectic strategies to improve reasoning beyond English, but these methods are related to specific language that is not always optimal for reasoning. |
| Approach: | They propose a modular approach that instructs models to structure reasoning passages in a different problem space and then self-refines their capabilities to deliver step-wise reasoning passage. |
| Outcome: | The proposed approach achieves significant improvements in multilingual reasoning of various models and task, with improved reasoning consistency across languages. |