Dataset Viewer
Auto-converted to Parquet Duplicate
name_test
string
round
int64
timestamp
string
summary
dict
stats
dict
results
list
test1
8
2025-12-27T19:26:16.503710
{ "total_judged": 125, "win_rate": 0.088, "margin": 0.024, "avg_gen_time": 19.496174279848734 }
{ "total": 128, "wins": 11, "draws": 106, "losses": 8, "skipped": 0, "gen_failed": 0, "render_failed": 0, "judge_failed": 3 }
[ { "task_id": "014635059549225c114d06bec52227b2c72ff0cb4dd8fd7e16cebfb596b8df45", "status": "success", "gen_time": 0, "winner": "left", "ours_penalty": 4, "leader_penalty": 6, "issues": "Direct pass: First model has color and transparency differences; second model has style (solid vs semi...

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

📊 Task Overview

Metric Value
Total tasks 128
Total judged 125
Skipped 0
Generation failed 0
Render failed 0
Judge failed 3
🏁 Results
Outcome Count
------- ----:
Wins 11
Draws 106
Losses 8
📈 Performance Metrics

Win Rate: 8.8%

Margin: 2.4%

Average Generation Time: 19.50 seconds

📝 Notes

No tasks were skipped or failed during generation/rendering.

A small number of tasks (3) failed during judging.

The majority of evaluated tasks resulted in draws, indicating close or balanced performance.

Downloads last month
18