American Legal Corpus Sentences Labeled for Citation Worthiness

This dataset benchmarks citation worthiness detection in the American Legal Corpus. It contains sentences from legal case files, each labeled with a binary integer indicating whether the sentence is citation-worthy. The dataset was created by Vidhaan and last updated in October 2023.

Use Cases

• Train a binary classifier using the Sentence and Label fields to predict citation worthiness.
• Analyze the relationship between Sentence Number and Label to identify positional patterns of citation-worthy content.
• Benchmark sequence models on the Sentence field for the task of legal text classification.
• Investigate citation patterns across different legal cases using the File Name and Label fields.

Strengths

• Dataset is focused on the American Legal Corpus, providing region-specific legal text.
• Includes a binary Label field for supervised learning tasks.
• Last updated in October 2023, indicating recent maintenance.

Limitations

• The total number of rows and columns is unknown, making it difficult to assess scale and feature richness.
• Sample data is unavailable, preventing preview of data structure and quality.
• The specific time range and geographic scope beyond 'US' are not detailed.

Provenance

Source

Hugging Face dataset uploaded by Vidhaan.

Collection Method

Sentences extracted from American Legal Corpus case files, with citation spans removed and labeled for citation worthiness.

Freshness

2023-10-23

Geography

United States (US)