Venues
What gets a paper rejected at NeurIPS, ICLR, or COLM? Instead of recycling folklore, we cluster the weaknesses reviewers actually flag in their public reviews and present them in a browsable way. Open science on open data.
ICLR 2024
2024Flagship venue for representation learning. Single round, fully transparent review cycle.
ICLR 2025
2025ICLR 2025 edition. Same format as 2024, an even larger dataset: the API exposes 11,672 submissions and 5,019 rejections. Lets us see how reviewer concerns evolve year over year.
NeurIPS 2024
2024The largest ML venue. OpenReview only exposes submissions that reached the decision stage; what you see here is the tail of the process.
TMLR
2024Transactions on Machine Learning Research. Journal-style continuous review with no deadline. Different incentives for authors and reviewers: does it translate into different rejection patterns?
Year-over-year: ICLR 2024 vs 2025
Same venue, two consecutive editions. Did what reviewers complain about change between 2024 and 2025?
Raw evidence
Every cluster includes anonymous exemplar quotes and its most representative terms. We don't ask you to trust us; we show the evidence.
Practical takeaways
Every pattern ships with an actionable note. What you'd change in your paper if you knew the pattern was there.
Reproducible
Open pipeline: TF-IDF + LSA + KMeans over OpenReview public data. Documented in OC-TR-2026-007.
How is this built?
Methodology, data sources, and limitations. Designed for reproducibility and auditing.