datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
FinRAGBench-V
FinRAGBench-V: A Benchmark for Multimodal RAG with Visual Citation in the Financial Domain 🤗 Code 📄 Paper
Overview
FinRAGBench-V is a comprehensive benchmark for visual retrieval-augmented generation (RAG) in finance, addressing the challenge that most existing financial RAG research focuses predominantly on text while overlooking rich visual content in financial documents. By integrating multimodal data and providing visual citation, FinRAGBench-V ensures traceability… See the full description on the dataset page: https://huggingface.co/datasets/zhaosuifeng/FinRAGBench-V.Fin-RATE
📝 Fin-RATE: Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings
Fin-RATE is a real-world benchmark to evaluate large language models (LLMs) on professional-grade reasoning over U.S. SEC filings.
It targets financial analyst workflows that demand:
📄 Long-context understanding
⏱️ Cross-year tracking
🏢 Cross-company comparison
📊 Structured diagnosis of model failures
📘 [Paper (arXiv link TBD)] | 🤗 Dataset
⬇️ SEC-based QA benchmark with 7,500… See the full description on the dataset page: https://huggingface.co/datasets/GGLabYale/Fin-RATE.finra-brokercheck-scraper
FINRA BrokerCheck Scraper · Advisors, Firms & Disclosures
Scrape financial advisors, firm affiliations, CRDs, registration scope, and disclosure histories directly from FINRA BrokerCheck API into clean dataset rows.
Rows in this dataset
1,430
Fields
22
Collector runs behind it
50
Most recent observation
2026-08-03
What this is
Every row here was returned by a real run of a public collector. Nothing is generated from a
template over a… See the full description on the dataset page: https://huggingface.co/datasets/reapxdev/finra-brokercheck-scraper.FinRAG
FinRAG
Dataset Description
This dataset contains 12,500 financial reasoning questions based on real-world financial documents, earnings reports, and financial tables. Each question is accompanied by a correct answer and four carefully crafted distractor answers, making it suitable for multiple-choice question answering tasks and assessing financial numerical reasoning capabilities.
Dataset Summary
Total Examples: 12,500
Format: Multiple-choice questions with 5… See the full description on the dataset page: https://huggingface.co/datasets/trismik/FinRAG.FinRAG-GRPO
FinRAG-GRPO Preference Dataset
A Chinese-language preference dataset for training Reasoning Reward Models (ReasRM) via GRPO-based reinforcement learning.
🚧 This dataset is actively maintained and will be expanded with additional domains and languages over time.
Dataset Summary
This dataset contains pairwise preference samples designed to train a reward model that reasons before judging — the model generates an evaluation rationale before outputting a preference label… See the full description on the dataset page: https://huggingface.co/datasets/SamWang0405/FinRAG-GRPO.Fin-RATE
📝 Fin-RATE: Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings
Fin-RATE is a real-world benchmark to evaluate large language models (LLMs) on professional-grade reasoning over U.S. SEC filings.
It targets financial analyst workflows that demand:
📄 Long-context understanding
⏱️ Cross-year tracking
🏢 Cross-company comparison
📊 Structured diagnosis of model failures
📘 [Paper (arXiv link TBD)] | 🤗 Dataset
⬇️ SEC-based QA benchmark with 7,500… See the full description on the dataset page: https://huggingface.co/datasets/idleengine/Fin-RATE.finRAG
finRAG Datasets
This is the official Huggingface repo of the finRAG datasets published by parsee.ai.
More detailed information about the 3 datasets and methodology can be found in the sub-directories for the individual datasets.
We wanted to investigate how good the current state of the art (M)LLMs are at solving the relatively simple problem of extracting revenue figures from publicly available financial reports. To test this, we created 3 different datasets, all based on the same… See the full description on the dataset page: https://huggingface.co/datasets/parsee-ai/finRAG.RAPID_dummy_pick_place_NAILONG_finray_0127This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "dummy_follower",
"total_episodes": 50,
"total_frames": 12915,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:50"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/inz/RAPID_dummy_pick_place_NAILONG_finray_0127.finrag-eval
FinRAG-Eval
147 questions over five S&P-500 10-K filings (563 printed pages, 1,679 chunks),
where every answerable question carries character-offset gold spans into the
canonical markdown — not just a gold string.
The point of this dataset is not its size. It is that it survived an audit
trail instead of a vibe check, and the trail is published with it.
Why another financial QA set
Most synthetic eval sets are generated once and trusted. This one was generated… See the full description on the dataset page: https://huggingface.co/datasets/ChihebLovesAi/finrag-eval.finra-firms
FINRA BrokerCheck firms (broker-dealers + registered investment advisers)
13,285 broker-dealers and registered investment advisers (RIAs) indexed by FINRA. BrokerCheck is presented as a search-UI-only product at brokercheck.finra.org; the backing JSON API isn't documented anywhere. The disclosure_fl=Y flag is the regulatory red-flag tripwire that KYC vendors charge $50k+/yr for. 922 firms flagged.
Live API
This dataset is served via a live REST API at… See the full description on the dataset page: https://huggingface.co/datasets/emperor-mew/finra-firms.FinRAD_Financial_Readability_Assessment_Dataset
FinRAD: Financial Readability Assessment Dataset - 13,000+ Definitions of Financial Terms for Measuring Readability
This repository contains the dataset mentioned in the paper: FinRAD: Financial Readability Assessment Dataset - 13,000+ Definitions of Financial Terms for Measuring Readability (presented at The Financial Narrative Processing Workshop colocated with LREC-2022, Marseille, France).
In addition to this, data collection & cleaning scripts, embedding extraction & model… See the full description on the dataset page: https://huggingface.co/datasets/sohomghosh/FinRAD_Financial_Readability_Assessment_Dataset.RAPID_dummy_pick_place_green_cube_finray_0127This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "dummy_follower",
"total_episodes": 50,
"total_frames": 13962,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:50"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/inz/RAPID_dummy_pick_place_green_cube_finray_0127.RAPID_dummy_pick_place_cola_bottle_finray_0127This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "dummy_follower",
"total_episodes": 5,
"total_frames": 1789,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:5"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/inz/RAPID_dummy_pick_place_cola_bottle_finray_0127.FinRAD_samplefin_ragKgxitxIdjxbdjdIddiuvuYuhiAsjuhJsisjshdjLintingLokmHaishzTuykaahrIschxhPopoOanjixPqowknsgfin-rag-bench