FMD — financial misinformation detection
Collection
FMDLlama (arXiv:2409.16452) data and the FMD 2025 challenge. • 3 items • Updated
Request access to FMD — financial misinformation detection
FMD — financial misinformation detection 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 FMDLlama paper in any resulting publication.
Log in or Sign Up to review the conditions and access this dataset content.
FMD — financial misinformation detection
📄 Paper · 🌐 The Fin AI
Part of FMDLlama — FMDLlama: Financial Misinformation Detection based on Large Language Models (arXiv:2409.16452). Formerly
TheFinAI/FMD_test(the old name redirects here).
Determine whether a financial claim is false, true or not supported by enough information, given context. The authors have not yet specified a license for this dataset; treat it as research-only until clarified, and respect the terms of the original sources.
| Task | claim verification (False / True / NEI) |
| Language | en |
| License | other (unspecified) |
Quick Start
from datasets import load_dataset
ds = load_dataset("TheFinAI/en-fmd", split="test")
print(ds[0])
| Split | Rows |
|---|---|
train |
4,462 |
val |
991 |
test |
2,804 |
Columns: id, query, answer, choices, gold
Citation
@misc{liu2025fmdllamafinancialmisinformationdetection,
title={FMDLlama: Financial Misinformation Detection based on Large Language Models},
author={Zhiwei Liu and Xin Zhang and Kailai Yang and Qianqian Xie and Jimin Huang and Sophia Ananiadou},
year={2025},
eprint={2409.16452},
archivePrefix={arXiv},
primaryClass={cs.CL},
doi={https://doi.org/10.1145/3701716.3715599},
url={https://arxiv.org/abs/2409.16452},
}
- Downloads last month
- 35