PapersWithELO

ICLR 2024 Authors

730 papers · aggregate stats only

Submission volume vs. outcome

Each point is a paper: x = submissions by its most prolific author, y = fused percentile. Correlation r = 0.151 slightly positive: higher submission volume associates with higher-ranked papers.

Submission-count distribution

Papers by their most prolific author's submission count. 193 papers have an author with ≥ 8 submissions.

1
124
2
128
3
81
4
56
5
59
6
54
7
35
8
27
9
25
10+
141

The mass-submission adjustment

Author credibility is computed from submission count: credibility = 1 / (1 + 0.15 · n_submissions)

When the leaderboard toggle is ON, a paper's fused percentile is shifted by a delta derived from its most prolific author's credibility — a transparent, reversible penalty against submission flooding. It is off by default, shown as an explicit delta per paper, and never hidden inside the score.

We publish aggregate statistics only. No per-author leaderboards, no names — the goal is to correct an incentive distortion, not to shame anyone.