百科.dev
全部条目在线工具AI 编程趋势榜开源项目技术资讯提交条目
登录
< 返回工具列表
M

MAST

> 编程语言
开源

我们的博客文章: https://sites.google.com/berkeley.edu/mast/ 此仓库包含了论文"为什么多代理系统会出现不一致的行为?"的代码和数据。

400 stars0 点赞0 次浏览
访问官网GitHub

工具介绍

我们的博客文章: https://sites.google.com/berkeley.edu/mast/ 此仓库包含了论文"为什么多代理系统会出现不一致的行为?"的代码和数据。

Our blogpost: https://sites.google.com/berkeley.edu/mast/

This repository contains the code and the data for the paper "Why Do Multi-Agent Systems Fail?" https://arxiv.org/pdf/2503.13657

In this paper, we present the first comprehensive study of MAS challenges: MAST (Multi-Agent Systems Failure Taxonomy).

News! We just released our dataset with over 1K annotated MAS traces https://arxiv.org/pdf/2503.13657v2

For LLM-as-a-Judge annotated traces:

from huggingface_hub import hf_hub_download
import pandas as pd
import json

REPO_ID = "mcemri/MAD"
FILENAME = "MAD_full_dataset.json"

file_path =  hf_hub_download(repo_id=REPO_ID, filename=FILENAME, repo_type="dataset")
with open(file_path, "r") as f:
    data = json.load(f)

print(f"Loaded {len(data)} records (full dataset).")

For human annotated traces:

FILENAME = "MAD_human_labelled_dataset.json"

file_path =  hf_hub_download(repo_id=REPO_ID, filename=FILENAME, repo_type="dataset")
with open(file_path, "r") as f:
    data = json.load(f)

print(f"Loaded {len(data)} records (human labelled).")

If you find this work useful, please cite it as follows:

bibtex
@article{cemri2025multi,
  title={Why Do Multi-Agent LLM Systems Fail?},
  author={Cemri, Mert and Pan, Melissa Z and Yang, Shuyi and Agrawal, Lakshya A and Chopra, Bhavya and Tiwari, Rishabh and Keutzer, Kurt and Parameswaran, Aditya and Klein, Dan and Ramchandran, Kannan and others},
  journal={arXiv preprint arXiv:2503.13657},
  year={2025}
}

Issues· 14 开放

查看全部 Issues在 GitHub 打开

暂无开放 Issues,或尚未同步最近议题。

> 标签

Python

暂无评论,来聊聊你的看法吧

> 工具信息

发布日期2026年8月1日
最后更新2026年9月17日
分类编程语言
定价开源

> 相关工具

T
TypeScript
JavaScript 的超集,为前端与全栈提供静态类型
P
Python
通用编程语言,广泛用于 Web、数据与 AI
G
Go
Google 推出的简洁高效系统语言