arXiv cs.CL· Shangbin Feng, Yuyang Bai, Ziyuan Yang, Yike Wang, Zhaoxuan Tan, Jiajie Yan, Zhenyu Lei, Wenxuan Ding, Weijia Shi, Haojin Wang, Zhenting Qi, Yuru Jiang, Heng Wang, Chengsong Huang, Yu Fei, Jihan Yao, Yilun Du, Luke Zettlemoyer, Yejin Choi, Yulia Tsvetkov·· 2 天前AI 评分40
MoCo：一站式模型协作研究 Python 库
MoCo: A One-Stop Shop for Model Collaboration Research
AI 导读
研究者推出 MoCo，一个用于大规模执行、评测和比较模型协作算法的一站式 Python 库，涵盖 26 种模型协作方法与 25 个评测数据集。实验显示，多数协作策略在平均 61.0% 的（模型、数据）设定中优于无协作模型，最优方法提升幅度最高达 25.8%。
正文 · 原文
Authors:Shangbin Feng, Yuyang Bai, Ziyuan Yang, Yike Wang, Zhaoxuan Tan, Jiajie Yan, Zhenyu Lei, Wenxuan Ding, Weijia Shi, Haojin Wang, Zhenting Qi, Yuru Jiang, Heng Wang, Chengsong Huang, Yu Fei, Jihan Yao, Yilun Du, Luke Zettlemoyer, Yejin Choi, Yulia Tsvetkov
Abstract:Advancing beyond single monolithic language models (LMs), recent research increasingly recognizes the importance of model collaboration, where multiple LMs collaborate, compose, and complement each other. Existing research on this topic has mostly been disparate and disconnected, from different research communities, and lacks rigorous comparison. To consolidate existing research and establish model collaboration as a school of thought, we present MoCo: a one-stop Python library of executing, benchmarking, and comparing model collaboration algorithms at scale. MoCo features 26 model collaboration methods, spanning diverse levels of cross-model information exchange such as routing, text, logit, and model parameters. MoCo integrates 25 evaluation datasets spanning reasoning, QA, code, safety, and more, while users could flexibly bring their own data. Extensive experiments with MoCo demonstrate that most collaboration strategies outperform models without collaboration in 61.0% of (model, data) settings on average, with the most effective methods outperforming by up to 25.8%. We further analyze the scaling of model collaboration strategies, the training/inference efficiency of diverse methods, highlight that the collaborative system solves problems where single LMs struggle, and discuss future work in model collaboration, all made possible by MoCo. We envision MoCo as a valuable toolkit to facilitate and turbocharge the quest for an open, modular, decentralized, and collaborative AI future.
| Comments: | Moco is available at this https URL |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2601.21257 [cs.CL] |
| (or arXiv:2601.21257v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2601.21257 arXiv-issued DOI via DataCite |
Submission history
From: Shangbin Feng [view email]
[v1]
Thu, 29 Jan 2026 04:36:52 UTC (413 KB)
[v2]
Sun, 19 Apr 2026 21:04:27 UTC (413 KB)
[v3]
Wed, 14 Oct 2026 22:11:07 UTC (413 KB)
来源：arXiv cs.CL · arxiv.org