Posts from the Inkhaven writing residency, ranked by a Bradley-Terry model fit on LLM pairwise judgments. Scores are expressed as standard deviations above the dataset mean, with bootstrap 95% confidence intervals.
| # | Score (σ, mean ± SE) | Author | Title | Date |
|---|
| # | Avg score (σ) | Posts | Author | Best post |
|---|
| # | Score (σ, mean ± SE) | Author | Title |
|---|
Scores come from a Bradley-Terry model fit on pairwise LLM rankings of all Inkhaven posts. Each post's σ-score is its BT log-strength divided by the dataset standard deviation; the ± uncertainty is one bootstrap standard error across resampled ranking groups.