ResearchBenchmarksDialog Response Generation on FEWSHOTWOZ Taxi 1.0 (test)Follow19.7BLEUSC-GPT11.286413.470715.65517.8393Feb 27, 2020Evaluation ResultsMethodMethodLinksBLEUERRSC-GPTbackbone=GPT2-Medium,...backbone=GPT2-Medium, number_of_parameters=345M2020.0219.73.57GPT-2backbone=GPT2-Medium,...backbone=GPT2-Medium, number_of_parameters=345M2020.0216.279.52SC-LSTMmodel_type=Semanticall...model_type=Semantically Conditioned LSTM2020.0211.6161.45