Question Answering on HotpotQA (Accuracy, Recall)
55.321AccuracyRAPTOR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RAPTORLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token, Clustering method=K-means2025.03 | 55.321 | 62.424 | |
| VanillaRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 50.783 | 57.745 | |
| HippoRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 50.324 | 58.86 | |
| FastGraphRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 43.193 | 51.007 | |
| HLightRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 35.647 | 43.334 | |
| ZeroShotLLM=Llama-3-8B2025.03 | 35.467 | 42.407 | |
| LLightRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 34.144 | 41.811 | |
| DALKLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 33.252 | 47.232 | |
| LGraphRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 33.063 | 42.691 | |
| GLightRAGLLM=Llama-3-8B, Embedding model=BGE-M3, top-k=4, chunk size=1,200-token2025.03 | 25.581 | 33.297 |