Question Answering on NQ (ASR, RSR, PPL)
81ASRAuthChain
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AuthChainModel=GPT-4o2025.05 | 81 | — | — | |
| AuthChainModel=Llama3-70B2025.05 | 81 | — | — | |
| AuthChainModel=Deepseek-V3-03242025.05 | 80 | — | — | |
| AuthChainModel=Llama3-8B2025.05 | 76 | — | — | |
| AuthChainModel=GPT-42025.05 | 75 | — | — | |
| AuthChainModel=GPT-3.52025.05 | 74 | — | — | |
| CoEModel=Deepseek-V3-03242025.05 | 68 | — | — | |
| PRAGModel=Deepseek-V3-03242025.05 | 67 | — | — | |
| CoEModel=Llama3-70B2025.05 | 66 | — | — | |
| HRAGModel=GPT-42025.05 | 64 | — | — | |
| CoEModel=GPT-4o2025.05 | 64 | — | — | |
| PRAGModel=Llama3-70B2025.05 | 64 | — | — | |
| CoEModel=Llama3-8B2025.05 | 63 | — | — | |
| AuthorityModel=Llama3-70B2025.05 | 61 | — | — | |
| AuthorityModel=Deepseek-V3-03242025.05 | 60 | — | — | |
| AuthorityModel=Llama3-8B2025.05 | 59 | — | — | |
| PRAGModel=Llama3-8B2025.05 | 57 | — | — | |
| CoEModel=GPT-3.52025.05 | 55 | — | — | |
| CoEModel=GPT-42025.05 | 55 | — | — | |
| PRAGModel=GPT-3.52025.05 | 54 | — | — | |
| HRAGModel=Deepseek-V3-03242025.05 | 54 | — | — | |
| AuthorityModel=GPT-3.52025.05 | 52 | — | — | |
| PRAGModel=GPT-42025.05 | 51 | — | — | |
| AuthorityModel=GPT-42025.05 | 51 | — | — | |
| AuthorityModel=GPT-4o2025.05 | 48 | — | — | |
| HRAGModel=Llama3-8B2025.05 | 48 | — | — | |
| PRAGModel=GPT-4o2025.05 | 44 | — | — | |
| HRAGModel=GPT-3.52025.05 | 40 | — | — | |
| HRAGModel=GPT-4o2025.05 | 38 | — | — | |
| HRAGModel=Llama3-70B2025.05 | 18 | — | — | |
| AuthChainModel=-2025.05 | — | 94 | 26.4 | |
| AuthorityModel=-2025.05 | — | 65 | 49.5 | |
| CoEModel=-2025.05 | — | 93 | 28.5 | |
| HRAGModel=-2025.05 | — | 87 | 398 | |
| PRAGModel=-2025.05 | — | 97 | 73.8 |