Long-context Question Answering on LongBench-Cite Average
77.6C ScoreClaude-3-sonnet
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Claude-3-sonnetCitation Strategy=LAC-S2024.09 | 77.6 | 78.3 | 99 | |
| GLM-4Citation Strategy=LAC-S2024.09 | 73.7 | 77.2 | 95 | |
| Mistral-Large-InstructCitation Strategy=LAC-S2024.09 | 73.6 | 76.4 | 96 | |
| LongCite-8BCitation Strategy=LAC-S2024.09 | 71.7 | 67.6 | 107 | |
| LongCite-9BCitation Strategy=LAC-S2024.09 | 70.4 | 65.6 | 109 | |
| GPT-4oCitation Strategy=LAC-S2024.09 | 69.4 | 78.2 | 88 | |
| GLM-4-9B-chatCitation Strategy=LAC-S2024.09 | 62.3 | 70.8 | 88 | |
| Llama-3.1-70B-InstructCitation Strategy=LAC-S2024.09 | 62 | 65.5 | 95 | |
| Llama-3.1-8B-InstructCitation Strategy=LAC-S2024.09 | 52.1 | 60.2 | 86 |