Multi-hop Question Answering on HOTPOTQA (EM, P, R, F1, BSc, HSc)
61EMKCS
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| KCSLLM=LLAMA3.1, Diversifying Method=+KCS2025.08 | 61 | 73.59 | 72.58 | 72.04 | 66.51 | 77.75 | |
| RASTLLM=LLAMA3.1, Diversifying Method=+RAST2025.08 | 55.5 | 67.8 | 68.33 | 66.49 | 59.93 | 71.4 | |
| Composition SamplingLLM=LLAMA3.1, Diversifying Method=+Composition2025.08 | 55 | 67.17 | 67.48 | 66.09 | 61.82 | 72.65 | |
| OriginalLLM=LLAMA3.1, Diversifying Method=ORI2025.08 | 54 | 66.07 | 65.29 | 64.46 | 58.74 | 69.05 | |
| GPT4LLM=LLAMA3.1, Diversifying Method=+GPT42025.08 | 49.5 | 60.81 | 61.45 | 59.75 | 52.43 | 65.55 | |
| KCSLLM=QWEN2.5, Diversifying Method=+KCS2025.08 | 45 | 60.49 | 58.54 | 57.4 | 51.56 | 67.55 | |
| Composition SamplingLLM=QWEN2.5, Diversifying Method=+Composition2025.08 | 40.5 | 55.7 | 52.29 | 52.02 | 46.43 | 59.7 | |
| OriginalLLM=QWEN2.5, Diversifying Method=ORI2025.08 | 39.5 | 54.9 | 51.46 | 51.36 | 47.16 | 60.4 | |
| RASTLLM=QWEN2.5, Diversifying Method=+RAST2025.08 | 34.5 | 49.57 | 46.66 | 45.96 | 39.86 | 54.85 | |
| GPT-4LLM=QWEN2.5, Diversifying Method=+GPT-42025.08 | 32.5 | 45.93 | 42.51 | 42.33 | 35.14 | 50.75 |