Summarization Faithfulness Evaluation on FeedSum
96Consistency ScoreGemini
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GeminiEvaluation framework=FineSurE2026.04 | 96 | 89 | 93 | 93 | |
| GPT4ominiEvaluation framework=FineSurE2026.04 | 95 | 93 | 93 | 94 | |
| GPT4-turboEvaluation framework=FineSurE2026.04 | 95 | 92 | 92 | 93 | |
| GPT4oEvaluation framework=FineSurE2026.04 | 94 | 91 | 95 | 93 | |
| MistralEvaluation framework=FineSurE2026.04 | 91 | 89 | 90 | 90 | |
| QwenEvaluation framework=FineSurE2026.04 | 88 | 83 | 85 | 85 | |
| Qwen*Evaluation framework=FineSurE2026.04 | 87 | 83 | 85 | 85 | |
| SummLLaMAEvaluation framework=FineSurE2026.04 | 85 | 85 | 84 | 85 | |
| LLaMA*Evaluation framework=FineSurE2026.04 | 85 | 80 | 80 | 82 | |
| LLaMAEvaluation framework=FineSurE2026.04 | 83 | 83 | 80 | 82 | |
| Mistral*Evaluation framework=FineSurE2026.04 | 81 | 82 | 80 | 81 |