Sentence Compression Pairs for Semantic Similarity

Aggregating pairs of equivalent sentences for sentence compression tasks. It is categorized as containing between 100,000 and 1 million entries (100K<n<1M) and is in English. The dataset was released by embedding-data and last updated in August 2022.

Use Cases

• Train a semantic similarity classification model to identify equivalent sentence pairs.
• Fine-tune a text generation model for sentence compression using the provided sentence pairs.
• Analyze linguistic patterns in English sentence compression within the US region context.

Strengths

• Contains between 100,000 and 1 million data points, as indicated by its size category.
• Focuses on a specific task: sentence compression and semantic similarity for English text.
• Includes metadata tags specifying the language (en) and region (us) for context.

Limitations

• Specific column definitions, sample data, and exact row count are unavailable, limiting initial analysis.
• The dataset's creation method and original source are not detailed, affecting reproducibility.
• The data is from 2022 or earlier, which may not reflect current linguistic usage.

Provenance

Source

embedding-data

Freshness

Last updated on 2022-08-02.

Geography

United States (inferred from 'us' region tag)