Multi-hop Knowledge Editing on MQUAKE-CF-3K All edited
45.87AccuracyPokeMQA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| PokeMQABackbone=GPT-3.5-turbo-instruct, Model Size=Undisclosed, Prompting Style=Question decomposition2023.12 | 45.87 | 39.77 | |
| MeLLoBackbone=GPT-3.5-turbo-instruct, Model Size=Undisclosed, Prompting Style=Question decomposition2023.12 | 35.27 | 25.3 | |
| PokeMQABackbone=LLaMa-2, Model Size=7B, Prompting Style=Question decomposition2023.12 | 32.83 | 23.87 | |
| PokeMQABackbone=Vicuna, Model Size=7B, Prompting Style=Question decomposition2023.12 | 31.63 | 25.3 | |
| MeLLoBackbone=LLaMa-2, Model Size=7B, Prompting Style=Question decomposition2023.12 | 17.33 | 9.9 | |
| MeLLoBackbone=Vicuna, Model Size=7B, Prompting Style=Question decomposition2023.12 | 10.9 | 6.7 | |
| MEMITBackbone=LLaMa-2, Model Size=7B, Prompting Style=Chain-of-thought2023.12 | 5.57 | — | |
| ROMEBackbone=LLaMa-2, Model Size=7B, Prompting Style=Question decomposition2023.12 | 3.63 | 0.1 | |
| ROMEBackbone=LLaMa-2, Model Size=7B, Prompting Style=Chain-of-thought2023.12 | 2.77 | — | |
| MEMITBackbone=LLaMa-2, Model Size=7B, Prompting Style=Question decomposition2023.12 | 2.3 | 0.37 |