Relevance Evaluation on SummEval
49.5AccuracyCollabEval
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| CollabEval2026.03 | 49.5 | 2.073 | 87.8 | 12 | 0.5 | 0 | 31.9 | 68.1 | |
| SonnetMode=Single-LLM2026.03 | 47.7 | 1 | 85.5 | 13.7 | 1.6 | 0 | 29.7 | 70.3 | |
| HaikuMode=Single-LLM2026.03 | 47.6 | 1 | 84.9 | 14.7 | 1.1 | 0 | 30.2 | 69.8 | |
| Llama3Mode=Single-LLM2026.03 | 22.8 | 1 | 76.7 | 23.3 | 0 | 0 | 100 | 0 |