Soccer Refereeing Explanation Quality on SoccerRefBench Overall 1.0 (test)
3.43Referee 1 ScoreSoccerRef-Agents
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SoccerRef-AgentsEvaluation Protocol=Human Evaluation2026.04 | 3.43 | 3.81 | 3.7 | 3.65 | |
| GPT-4oEvaluation Protocol=Human Evaluation2026.04 | 3.32 | 3.33 | 3.41 | 3.36 |