Long-context Question Answering on Dureader
81C ScoreGPT-4o
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GPT-4oCitation Strategy=LAC-S2024.09 | 81 | 83.3 | 97 | |
| Mistral-Large-InstructCitation Strategy=LAC-S2024.09 | 79 | 83.3 | 95 | |
| GLM-4Citation Strategy=LAC-S2024.09 | 76 | 75.8 | 100 | |
| Claude-3-sonnetCitation Strategy=LAC-S2024.09 | 75.8 | 80.3 | 94 | |
| LongCite-9BCitation Strategy=LAC-S2024.09 | 69 | 66.3 | 104 | |
| LongCite-8BCitation Strategy=LAC-S2024.09 | 68.5 | 62.3 | 110 | |
| GLM-4-9B-chatCitation Strategy=LAC-S2024.09 | 49.3 | 68.1 | 72 | |
| Llama-3.1-70B-InstructCitation Strategy=LAC-S2024.09 | 43.3 | 42.5 | 102 | |
| Llama-3.1-8B-InstructCitation Strategy=LAC-S2024.09 | 29.8 | 39.4 | 76 |