datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
BVA-AMA-Era-170KStructured
MatrixNorm BVA AMA-Era 170K
MatrixNorm BVA AMA-Era 170K is a normalized legal-AI dataset built from public
Board of Veterans' Appeals (BVA) decision records from the AMA era.
This dataset converts messy public veterans-law decisions into structured,
AI-ready legal records with extracted issues, outcomes, conditions, citations,
judges, appeal-lane signals, metadata, and reasoning fields.
Current release: ~181K rows — but ≈108K unique decisions (this file contains
duplicate rows;… See the full description on the dataset page: https://huggingface.co/datasets/williamTLmiller/BVA-AMA-Era-170KStructured.Bva_ama_decisions__structured_2019
BVA AMA Decisions — Structured (A-prefix)
Structured annotations over 5,990 U.S. Board of Veterans' Appeals (BVA)
decisions issued under the Appeals Modernization Act (AMA) — the modernized
appeals system that replaced the legacy process. Every record is an A-prefix
docket decision (A21……), so this is a 100% AMA corpus: the part of the
BVA record that the standard academic/legacy datasets do not cover.
Why this dataset is different
The widely-used BVA legal-NLP… See the full description on the dataset page: https://huggingface.co/datasets/williamTLmiller/Bva_ama_decisions__structured_2019.bva-decisions-structured-sample2019Present
BVA Structured Decisions (2019–2025)
Structured, issue-level records extracted from U.S. Board of Veterans' Appeals (BVA) decisions — each decision parsed into its issues, conditions, outcomes, citations, and reasoning, with per-document provenance and completeness flags. Built for training and evaluating legal-AI models on veterans' disability adjudication.
This is a 2900-decision sample, balanced across seven years (2019–2025, decisions/year), so it's representative of the… See the full description on the dataset page: https://huggingface.co/datasets/williamTLmiller/bva-decisions-structured-sample2019Present.