yev1 Author Tuned-on Benchmark - R-Judge571: leaderboard

yev1 Author Tuned-on Benchmark - R-Judge571 — Author BF16 letter-logit decision-model evaluation. Explicitly tuned-on benchmarks: failure analysis of these test sets informed training data; NOT held-out scores. Existing DecideBench headline already covered. Metric: F1 (%). Source: https://huggingface.co/choyiny/yev1-4b

Metric: F1 (%). Source: huggingface.co. 1 model tracked. Updated 16 Nov 2026.

Top models

#ModelScoreOverall rank
1yev1-4B80.8

Interactive version: theaggregate.ai/benchmark?slug=yev1-author-tuned-on-benchmark-r-judge571 · How It Works · Data refreshed daily, snapshot 2026-11-16.