LIBRA Old: Legacy Russian Long-Context NLP Benchmark

A legacy version of the LIBRA benchmark for analyzing long Russian text inputs, archived for reproducibility. The dataset was created by ai-forever and is no longer maintained, with a last recorded update on 2026-05-05. Users are directed to an updated version for cleaned data and extended context lengths.

Use Cases

• Benchmarking long-context NLP models based on the described Russian analysis tasks.
• Reproducing results from research papers that used the original LIBRA version.
• Studying the evolution of Russian language benchmarks by comparing legacy and updated datasets.

Strengths

• Archived for reproducibility, supporting verification of prior research.
• Created by a known organization, ai-forever, suggesting a research-oriented origin.

Limitations

• Explicitly marked as a legacy version and no longer maintained.
• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count, file formats, and column definitions are unknown, which may limit suitability assessment.

Provenance

Source

ai-forever

Collection Method

Likely constructed as a benchmark for Russian language analysis, as described in an associated paper.

Freshness

Last updated 2026-05-05 10:54:51; freshness should be verified.