Request access to FinTagging (original format)

FinTagging (original format) is released by The Fin AI for research. Access is granted automatically after you complete this short form.

By accessing this dataset you agree to its license and to cite the FinTagging paper in any resulting publication.

Log in or Sign Up to review the conditions and access this dataset content.

FinTagging (original format)

πŸ“„ Paper Β· πŸ€— Collection Β· πŸ’» Code Β· 🌐 The Fin AI

Part of FinTagging β€” FinTagging: Benchmarking LLMs for Extracting and Structuring Financial Information (arXiv:2505.20650). Formerly TheFinAI/FinTagging_Original (the old name redirects here).

FinTagging contexts with their tagged numeric entities in the original (non-BIO) format. A newer, larger version is available: en-fintagging-original-v2.

Task XBRL tagging
License CC BY 4.0
Language en

Quick Start

from datasets import load_dataset

ds = load_dataset("TheFinAI/en-fintagging-original", split="test")
print(ds[0])

Dataset Structure

Split Rows
test 6,599
Field Description
text Text or table context
numeric_entities Tagged numeric entities with their US-GAAP concepts
context_id Context id

Citation

@misc{wang2026fintaggingbenchmarkingllmsextracting,
      title={FinTagging: Benchmarking LLMs for Extracting and Structuring Financial Information}, 
      author={Yan Wang and Lingfei Qian and Xueqing Peng and Yang Ren and Keyi Wang and Yi Han and Dongji Feng and Fengran Mo and Shengyuan Lin and Qinchuan Zhang and Kaiwen He and Chenri Luo and Jianxing Chen and Junwei Wu and Chen Xu and Ziyang Xu and Jimin Huang and Guojun Xiong and Xiao-Yang Liu and Qianqian Xie and Jian-Yun Nie},
      year={2026},
      eprint={2505.20650},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.20650}, 
}
Downloads last month
46

Collection including TheFinAI/en-fintagging-original

Paper for TheFinAI/en-fintagging-original