Multi-paper Question Answering on NLP-MQA annotated (test)
78.3PrecisionCE-GOCD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CE-GOCDLLM=DeepSeek-V32026.01 | 78.3 | 82.11 | 80.14 | |
| CE-GOCDLLM=Qwen-plus2026.01 | 77.2 | 82.46 | 79.73 | |
| CE-GOCDLLM=GPT-42026.01 | 76.66 | 81.56 | 79.01 | |
| MindMapLLM=GPT-42026.01 | 76.28 | 73.86 | 74.96 | |
| LLM onlyLLM=Qwen-plus2026.01 | 72.31 | 74.61 | 73.38 | |
| PathRAGLLM=GPT-42026.01 | 72.2 | 77.89 | 74.82 | |
| LLM onlyLLM=DeepSeek-V32026.01 | 70.91 | 75.34 | 73.03 | |
| LLM onlyLLM=GPT-42026.01 | 69.85 | 70.39 | 70.06 | |
| BM25LLM=GPT-42026.01 | 68.9 | 70.97 | 69.89 | |
| Embedding retrievalLLM=GPT-42026.01 | 68.78 | 71.17 | 69.92 | |
| KAGLLM=GPT-42026.01 | 68.19 | 77.97 | 72.75 |