Not synced yet
OPENREVIEW ANALYTICS · CONFERENCE-WIDE

Understand your standing
in the review landscape.

Explore official review scores, rank submissions, and find exactly where your paper stands within the public ICLR 2027 distribution.

Active submissions ◫
—
Waiting for data
Scored papers ▤
—
Official reviews only
Mean paper rating ◎
—
Scale 1–4
Mean confidence ◇
—
Among scored reviews
DISTRIBUTION

Average paper ratings

Paper-level
1.0 · Strong reject4.0 · Strong accept
REVIEW LANDSCAPE

Score composition

Review-level
YOUR PAPER

Where does your average place you?

Compare your own rating against every scored paper in the selected cohort.

ENTER YOUR AVERAGE REVIEW SCORE
/ 4.00
1.002.003.004.00
ⓘ Comparison uses the chosen ranking metric and filters. Your paper is not included in the cohort unless it is already present as a submission.
YOUR RELATIVE STANDING All areas
Estimated rank—among scored papers
—Higher than you
—Lower than you
—Equal to you
Midrank percentile—
FULL LEADERBOARD

Paper rankings

Every scored paper, linked to the public OpenReview forum.

No records
RANKSUBMISSIONAREAMEANWEIGHTEDREVIEWSCONF.SPREAD
Load public OpenReview data to see the leaderboard.
REVIEWERS

Confidence and disagreement

Understand individual scores, score variance and reviewer confidence.

Review confidence

Scored reviews

Reviewer disagreement

Rated papers
Std. dev. 0Greater disagreement →
DATA SOURCE

Keep your results up to date

The website can load a published snapshot or fetch public OpenReview data into your browser.

↻

Load from OpenReview

Fetch every publicly available submission and official review. This can take several minutes; your browser must permit requests to the OpenReview API.

⇩

Import review-level CSV

Use a local export if direct API access is restricted. Required columns: paper_id,title,rating,confidence. Optional: review_id,number,area,status,url.

How rankings are calculated. Each paper's score is the arithmetic mean of its valid numeric official reviewer ratings. The optional weighted score is Σ(rating × confidence) / Σ(confidence), using reviews with positive confidence. Equal scores share a rank. Withdrawn, desk-rejected and rejected papers are excluded by default. All comparisons are unofficial and are not acceptance predictions.

No data source loaded.