Single-hop Question Answering on Loogle SD
45.1ScoreDCS
Evaluation results
| Method | Links | |
|---|---|---|
| DCSBackbone=Llama-3-8B-Instruct2025.06 | 45.1 | |
| DCSBackbone=Mistral-7B-Instruct2025.06 | 24.51 | |
| Llama-3-8B-InstructBackbone=Llama-3-8B-Instruct, Context Management=Baseline2025.06 | 21.25 | |
| MOICEBackbone=Mistral-7B-Instruct2025.06 | 20.81 | |
| DCSBackbone=Vicuna-7B2025.06 | 20.19 | |
| InfLLMBackbone=Llama-3-8B-Instruct2025.06 | 19.5 | |
| Mistral-7B-InstructBackbone=Mistral-7B-Instruct, Context Management=Baseline2025.06 | 19.13 | |
| StreamingBackbone=Llama-3-8B-Instruct2025.06 | 18.51 | |
| LM-infiniteBackbone=Llama-3-8B-Instruct2025.06 | 18.2 | |
| InfLLMBackbone=Mistral-7B-Instruct2025.06 | 18.15 | |
| MOICEBackbone=Vicuna-7B2025.06 | 14.63 | |
| Vicuna-7BBackbone=Vicuna-7B, Context Management=Baseline2025.06 | 14.11 | |
| LM-InfiniteBackbone=Vicuna-7B2025.06 | 13.68 | |
| StreamingBackbone=Vicuna-7B2025.06 | 13.11 | |
| LM-infiniteBackbone=Mistral-7B-Instruct2025.06 | 12.06 | |
| StreamingBackbone=Mistral-7B-Instruct2025.06 | 11.8 | |
| InfLLMBackbone=Vicuna-7B2025.06 | 11.29 |