new elo results

allenai · Jul 15, 2024 · dd82749 · dd82749
1 parent de573cf
commit dd82749
Show file tree

Hide file tree

Showing 12 changed files with 1,734 additions and 1,379 deletions.
diff --git a/leaderboard/data_dir/all_stat.json b/leaderboard/data_dir/all_stat.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.-1.json b/leaderboard/data_dir/all_stat_wildbench.-1.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.100.json b/leaderboard/data_dir/all_stat_wildbench.100.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.1000.json b/leaderboard/data_dir/all_stat_wildbench.1000.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.1500.json b/leaderboard/data_dir/all_stat_wildbench.1500.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.2000.json b/leaderboard/data_dir/all_stat_wildbench.2000.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.300.json b/leaderboard/data_dir/all_stat_wildbench.300.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.3000.json b/leaderboard/data_dir/all_stat_wildbench.3000.json
diff --git a/leaderboard/data_dir/all_stat_wildbench.500.json b/leaderboard/data_dir/all_stat_wildbench.500.json
diff --git a/leaderboard/data_dir/score.json b/leaderboard/data_dir/score.json
diff --git a/leaderboard/data_dir/wb_elo_results.json b/leaderboard/data_dir/wb_elo_results.json
diff --git a/leaderboard/show_eval.sh b/leaderboard/show_eval.sh
@@ -28,7 +28,8 @@ python leaderboard/data_dir/_create_tables.py score
 
 python leaderboard/data_dir/_merge_results.py
 
-margin=3;tie_margin=2;K=4;dynamic=True;interval=16
+# margin=3;tie_margin=2;K=4;dynamic=True;interval=16
+margin=3;tie_margin=2;K=4;dynamic=True;interval=100
 python -m leaderboard.wb_elo --K $K --margin $margin --tie_margin $tie_margin --num_rounds 100 --dynamic $dynamic --interval $interval --num_processes 4
 
 python leaderboard/data_dir/_merge_results.py