Coherence Evaluation on SummEval
55.9AccuracyMistral
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MistralMode=Single-LLM2026.03 | 55.9 | 1 | 93.8 | 5.4 | 2.8 | 0 | 24.4 | 75 | |
| CollabEval2026.03 | 48.2 | 2.181 | 79.6 | 18.2 | 7 | 0 | 10.2 | 89.8 | |
| SonnetMode=Single-LLM2026.03 | 46.9 | 1 | 65.8 | 25.2 | 19.8 | 0 | 10.4 | 89.6 | |
| CollabEval2026.03 | 40.4 | 2.343 | 77.8 | 20.8 | 1.5 | 0 | 63.3 | 36.7 | |
| HaikuMode=Single-LLM2026.03 | 38.9 | 1 | 76.9 | 22.4 | 0.8 | 0 | 63.4 | 36.6 | |
| SonnetMode=Single-LLM2026.03 | 37.4 | 1 | 71.4 | 23.9 | 4.9 | 0 | 66.4 | 33.6 | |
| Llama3Mode=Single-LLM2026.03 | 29.5 | 1 | 77 | 22 | 2.2 | 0 | 25.4 | 74.6 | |
| HaikuMode=Single-LLM2026.03 | 12.6 | 1 | 77.7 | 20.9 | 5.3 | 0 | 4.7 | 95.3 |