ResearchBenchmarksSequential Knowledge Editing on MQuAKEFollow97.4EfficacyD4S-3.89622.40248.774.998Oct 31, 2024Evaluation ResultsMethodMethodLinksEfficacyParaphrase ScoreSpecificityAverage ScoreD4SModel=GPTModel=GPT2024.1097.475.5421.8464.93D4SModel=LlamaModel=Llama2024.1085.372.6828.1662.05ROMEModel=LlamaModel=Llama2024.1076.8578.413.6752.97FTModel=LlamaModel=Llama2024.1041.4344.9328.2438.2FTModel=GPTModel=GPT2024.10176.2207.74ROMEModel=GPTModel=GPT2024.100000MEMITModel=GPTModel=GPT2024.100000MEMITModel=LlamaModel=Llama2024.100000