List your company

ResearchFormerly WizWand

Trajectory quality correlation analysis

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

Teacher-generated Reasoning Datasets Average across students

Rank-Surprisal Ratio

0.86Spearman Correlation (Abs)

13

Feb 26, 2026

Teacher-generated Reasoning Datasets Qwen-2.5-3B student

Rank-Surprisal Ratio

0.85Spearman Correlation (Abs)

13

Feb 26, 2026

Teacher-generated Reasoning Datasets Qwen-3-4B student

Rank-Surprisal Ratio

0.82Spearman Correlation (Absolute)

13

Feb 26, 2026

Teacher-generated Reasoning Datasets Qwen-2.5-7B student

Rank-Surprisal Ratio

0.92Spearman Correlation (Abs)

13

Feb 26, 2026

Teacher-generated Reasoning Datasets LLaMA-3.1-8B student

Rank-Surprisal Ratio

0.85Spearman Correlation (Abs)

13

Feb 26, 2026

Teacher-generated Reasoning Datasets Qwen-3-14B student

Rank-Surprisal Ratio

0.85Spearman Correlation (Abs)

13

Feb 26, 2026