BFCL-FC-robustness
收藏魔搭社区2026-01-06 更新2025-12-06 收录
下载链接:
https://modelscope.cn/datasets/ibm-research/BFCL-FC-robustness
下载链接
链接失效反馈官方服务:
资源简介:
# BFCL-FC-Robustness
This dataset is an extension of `BFCL-v3-multiple` dataset, and it includes:
- `BFCL_v3_multiple_p_quest.json`: a version with paraphrased user query
- `BFCL_v3_multiple_tool_enrichment.json`: a version with expanded agent toolkit
- `BFCL_v3_multiple_tool_enrichment_p_fdesc.json`: a version with expanded agent toolkit + empty function description
- `BFCL_v3_multiple_tool_enrichment_p_pdesc.json`: a version with expanded agent toolkit + empty parameter description
Used in our work on agentic function calling robustness described in this paper: https://aclanthology.org/2025.trustnlp-main.20/.
## Cite this Dataset
In case you use our dataset in your paper, please cite it by
```
@inproceedings{rabinovich-anaby-tavor-2025-robustness,
title = "On the Robustness of Agentic Function Calling",
author = "Rabinovich, Ella and Anaby Tavor, Ateret",
booktitle = "Proceedings of the 5th Workshop on Trustworthy NLP (TrustNLP 2025)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.trustnlp-main.20/",
doi = "10.18653/v1/2025.trustnlp-main.20",
year = "2025"
}
```
# BFCL-FC-Robustness
本数据集为`BFCL-v3-multiple`数据集的扩展版本,包含以下子文件:
- `BFCL_v3_multiple_p_quest.json`:包含经过释义改写的用户查询的版本
- `BFCL_v3_multiple_tool_enrichment.json`:包含扩展后的AI智能体(AI Agent)工具集的版本
- `BFCL_v3_multiple_tool_enrichment_p_fdesc.json`:包含扩展后的AI智能体(AI Agent)工具集且函数描述为空的版本
- `BFCL_v3_multiple_tool_enrichment_p_pdesc.json`:包含扩展后的AI智能体(AI Agent)工具集且参数描述为空的版本
本数据集应用于我们针对智能体函数调用鲁棒性的研究工作,相关研究详见论文:https://aclanthology.org/2025.trustnlp-main.20/。
## 数据集引用
若您在学术论文中使用本数据集,请按以下格式进行引用:
@inproceedings{rabinovich-anaby-tavor-2025-robustness,
title = "On the Robustness of Agentic Function Calling",
author = "Rabinovich, Ella and Anaby Tavor, Ateret",
booktitle = "第五届可信自然语言处理研讨会(TrustNLP 2025)论文集",
publisher = "国际计算语言学协会",
url = "https://aclanthology.org/2025.trustnlp-main.20/",
doi = "10.18653/v1/2025.trustnlp-main.20",
year = "2025"
}
提供机构:
maas
创建时间:
2025-10-12



