Dataset
kl3m-data-edgar-10-k
A component of the ALEA Institute's KL3M Data Project supplying training material cleared of copyright concerns. The dataset card is currently a placeholder, directing readers to the GitHub repo and project paper for full documentation.
Data preview
Dataset card
The kl3m-data-edgar-10-k dataset, published by the ALEA Institute as part of the KL3M Data Project, provides a copyright-clean corpus of U.S. Securities and Exchange Commission EDGAR 10-K filings formatted for large language model training. It is distributed in Parquet on Hugging Face under CC-BY terms alongside the project's broader pipeline.
What this dataset contains
The dataset holds text renditions of 10-K annual reports sourced from the SEC's EDGAR system, intended as copyright-clean pre-training material rather than for direct financial extraction. The single train split contains 1,194,043 rows, and the overall release is part of a 132-million-document corpus that the publisher verifies against copyright and contract terms. Tag metadata indicates a size between one million and ten million samples.
Structure and fields
Records are delivered in Parquet with the schema fields identifier (string), dataset (string), mime_type (string), and tokens (int64). The identifier column stores an S3-style path under data.kl3m.ai/documents/edgar/, the dataset column carries the source label edgar, mime_type is reported as text/markdown, and tokens holds the token count for the document. The dataset is compatible with the datasets, Dask, Polars, and mlcroissant libraries.
How analysts use it
Because the schema exposes only identifier, dataset, mime_type, and tokens, it is best understood as a text corpus for training or benchmarking language models on regulatory disclosures rather than a structured fundamentals feed. Researchers building NLP systems for filing summarization, risk-factor extraction, or compliance question-answering can use it as a vetted training source. Direct quantitative signal extraction for investment models would require re-parsing the underlying markdown, which falls outside what this dataset provides directly.
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.