OpenReview Raw: Peer Review Data from Major AI/ML Conferences

OpenReview Raw is a compilation of publicly available peer review data from the OpenReview platform. It includes reviews, official comments, meta-reviews, and decisions for 49,023 unique papers from venues like ICLR, NeurIPS, and EMNLP. The dataset was compiled by priorcomputers and was last updated on 2026-03 -13.

Use Cases

• Analyze review sentiment and decision patterns based on the included review text and decisions.
• Study the evolution of feedback in scientific communities based on data from multiple major conferences.
• Train models to predict paper acceptance based on features from review text and meta-reviews.
• Conduct meta-research on peer review quality and bias across different AI/ML venues.

Strengths

• Covers 49,023 unique papers, providing a substantial corpus for analysis.
• Includes multiple text components per submission: reviews, comments, and meta-reviews.
• Aggregates data from several major AI/ML conferences, offering cross-venue comparison potential.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown beyond the paper count, which may limit suitability assessment for specific tasks.
• Data may reflect temporal and source bias inherent to the specific conferences covered on OpenReview.

Provenance

Source

OpenReview platform

Collection Method

Compilation of publicly available data from OpenReview.

Time Range

null

Freshness

Last updated 2026-03-13 12:26:14; freshness should be verified.

Geography

null