Dataset
kl3m-data-edgar-agreements
Material contracts and agreements extracted from EDGAR filings by the KL3M project: debt, M&A, employment and license agreements, millions of documents in Parquet. Contract language is a niche but real input for M&A, financing and litigation-event signals.
Data preview
Dataset card
kl3m-data-edgar-agreements is a parquet text corpus of SEC EDGAR agreements distributed by the ALEA Institute as part of its KL3M copyright-clean training data project. The published split contains roughly 1.45 million rows of tokenized agreement text.
What this dataset contains
The release covers SEC EDGAR-filed agreement documents surfaced through the ALEA Institute KL3M Data Project, a broader 16-source corpus that the publisher describes as copyright- and contract-clean for large language model training. Documents are delivered in parquet format with a single train split of about 1.45 million rows. The dataset is hosted on S3, Hugging Face, and GitHub under CC-BY terms according to the publisher.
Structure and fields
Each row carries an identifier string pointing to the original S3 location of the document, a dataset label set to edgar-agreements, a mime_type field recorded as text/plain, and a tokens column storing an int64 sequence of pre-tokenized document representations. The repository is tagged for use with the Hugging Face datasets, Dask, mlcroissant, and polars libraries.
How analysts use it
Analysts can query the identifier, dataset, and mime_type fields to inventory and audit the underlying EDGAR agreement documents, while the tokens column is intended for downstream NLP rather than direct financial extraction. A natural research application is fine-tuning language models on contract language to support legal text classification, clause summarization, or agreement-structure analysis. Quantitative investment use cases are not the primary focus and are not directly supported by the schema.
Documentation & references 1
Related datasets
Browse moreParsed eligibility criteria from 13,229 ClinicalTrials.gov studies representing the candidate pool gathered by two first-stage retrievers during TREC Clinical Trials 2021–2023, formatted as typed entity-relation graphs for reranking models.
The clinical-trials-patient-graphs dataset is a small tabular collection of patient-level clinical records distributed in Parquet format by the 2001jdev publisher. It organizes entities and relations extracted from trial data into structured rows spanning 2021 through 2023.
This dataset is a parsed snapshot of ClinicalTrials.gov records, distributed in Parquet format on the Hugging Face Hub under an unspecified license. It contains roughly 52,000 trial entries drawn from public registry filings.
clinical-trials-trec-qrels is a tabular relevance-judgment file distributed by the 2001jdev user on the Hugging Face Hub. It maps clinical-trial topic identifiers to NCT registry IDs with graded relevance scores.
The clinical-trials-trec-topics dataset is a small tabular collection of clinical-trial topic records distributed in Parquet format on the Hugging Face Hub. It appears to be derived from TREC clinical-trial retrieval topics, with rows partitioned by year.
SEC EDGAR 10-K & 10-Q financial statements for 2,600+ NASDAQ companies.
Discussion
Questions or observations about this dataset? Join the conversation on X.
No comments yet. Start the conversation on X.