Lune

ACL2024Top-tier venue

Chat Vector: A Simple Approach to Equip LLMs with Instruction Following and Model Alignment in New Languages

Shih-Cheng Huang, Pin-Zu Li, Yu-Chi Hsu, Kuang-Ming Chen, Yu-Tung Lin, Shih-Kai Hsiao, Richard Tzong-Han Tsai, Hung-yi Lee

Abstract

Recently, the development of open-source large language models (LLMs) has advanced rapidly. Nevertheless, due to data constraints, the capabilities of most open-source LLMs are primarily focused on English. To address this issue, we introduce the concept of chat vector to equip pre-trained language models with instruction following and human value alignment via simple model arithmetic. The chat vector is derived by subtracting the weights of a pretrained base model (e.g. LLaMA2) from those of its corresponding chat model (e.g. LLaMA2chat). By simply adding the chat vector to a continual pre-trained model's weights, we can endow the model with chat capabilities in new languages without the need for further training. Our empirical studies demonstrate the superior efficacy of the chat vector from three different aspects: instruction following, toxicity mitigation, and multi-turn dialogue. Moreover, to showcase the adaptability of our approach, we extend our experiments to encompass various languages, base models, and chat vectors. The results underscore the chat vector's simplicity, effectiveness, and wide applicability, making it a compelling solution for efficiently enabling conversational capabilities in pre-trained language models. Our code is available at https: //github.com/aqweteddy/ChatVector .

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

13 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • LoRA: Low-Rank Adaptation of Large Language Models cites Language Models are Few-Shot Learners
  • Continual Model Merging without Data: Dual Projections for Balancing Stability and Plasticity cites Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
  • Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognition cites LoRA: Low-Rank Adaptation of Large Language Models
  • Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognition cites Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning
  • Direct Preference Optimization: Your Language Model is Secretly a Reward Model cites Language Models are Few-Shot Learners
  • Neural Incompatibility: The Unbridgeable Gap of Cross-Scale Parametric Knowledge Transfer in Large Language Models cites LoRA: Low-Rank Adaptation of Large Language Models
  • Neural Incompatibility: The Unbridgeable Gap of Cross-Scale Parametric Knowledge Transfer in Large Language Models cites Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
  • Neural Incompatibility: The Unbridgeable Gap of Cross-Scale Parametric Knowledge Transfer in Large Language Models cites Language Models are Few-Shot Learners
  • Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning cites LoRA: Low-Rank Adaptation of Large Language Models
  • Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning cites Language Models are Few-Shot Learners
  • Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded Updates cites LoRA: Low-Rank Adaptation of Large Language Models
  • Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded Updates cites Direct Preference Optimization: Your Language Model is Secretly a Reward Model
  • Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded Updates cites Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers13

  1. Continual Model Merging without Data: Dual Projections for Balancing Stability and PlasticityEnneng Yang, Anke Tang, Li Shen, Guibing Guo et al.NeurIPS 2025 · 13 citations
  2. Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech RecognitionHsuan Su, Hua Farn, Fan-Yun Sun, Shang-Tse Chen et al.EMNLP 2024 · 6 citations
  3. Personality Vector: Modulating Personality of Large Language Models by Model MergingSeungjong Sun, Seo Yeon Baek, Jang Hyun KimEMNLP 2025 · 5 citations
  4. Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded UpdatesAtsuki Yamaguchi, Terufumi Morishita, Aline Villavicencio, Nikolaos AletrasACL 2026 · 3 citations
  5. Neural Incompatibility: The Unbridgeable Gap of Cross-Scale Parametric Knowledge Transfer in Large Language ModelsYuqiao Tan, Shizhu He, Kang Liu, Jun ZhaoACL 2025 · 2 citations

Builds on12

  1. Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
  2. LoRA: Low-Rank Adaptation of Large Language ModelsEdward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu et al.ICLR 2022 · 18,833 citations
  3. Direct Preference Optimization: Your Language Model is Secretly a Reward ModelRafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D. Manning et al.NeurIPS 2023 · 10,924 citations
  4. Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context LearningHaokun Liu, Derek Tam, Mohammed Muqeeth, Jay Mohta et al.NeurIPS 2022 · 1,483 citations
  5. Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference timeMitchell Wortsman, Gabriel Ilharco, Samir Yitzhak Gadre, Rebecca Roelofs et al.ICML 2022 · 1,464 citations

Related papers

  1. FuseChat: Knowledge Fusion of Chat ModelsFanqi Wan, Longguang Zhong, Ziyi Yang, Ruijun Chen et al.EMNLP 2025 · 4 citations
  2. Emulated Disalignment: Safety Alignment for Large Language Models May Backfire!Zhanhui Zhou, Jie Liu, Zhichen Dong, Jiaheng Liu et al.ACL 2024
  3. Internal Value Alignment in Large Language Models through Controlled Value Vector ActivationHaoran Jin, Meng Li, Xiting Wang, Zhihao Xu et al.ACL 2025
  4. Efficient Model Development through Fine-tuning TransferPin-Jie Lin, Rishab Balasubramanian, Fengyuan Liu, Nikhil Kandpal et al.EMNLP 2025
  5. ChatMusician: Understanding and Generating Music Intrinsically with LLMRuibin Yuan, Hanfeng Lin, Yi Wang, Zeyue Tian et al.ACL Findings 2024 · 41 citations
Dusk over the sea between two cliffs drawn in fine vertical lines