Litbank-en: Standardized Coreference Resolution Dataset for Literature

litbank-en is a reformatted version of the original coreference resolution dataset for literature. The dataset is standardized to provide a unified document structure for cross-dataset comparison and multilingual experimentation. It was created by lattice-nlp and last updated on 2026-05-20.

Use Cases

• Benchmarking coreference resolution systems based on the standardized document structure.
• Conducting multilingual coreference experiments based on the dataset's interoperability design.
• Comparing coreference performance across datasets based on the unified formatting.
• Enabling reproducible evaluation settings for NLP pipelines based on the dataset's purpose.

Strengths

• Standardized and reformatted structure simplifies cross-dataset comparison.
• Designed for interoperability between NLP pipelines and reproducible evaluation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

lattice-nlp

Collection Method

Standardization and reformatting of the original litbank-en coreference resolution dataset.

Freshness

Last updated 2026-05-20 09:17:32; freshness should be verified.