ResearchBenchmarksDialogue Generation on DuRecDial ID (test)Follow3.17ProficiencyOurs2.40042.60022.82.9998May 12, 2026Evaluation ResultsMethodMethodLinksProficiencyCoherenceAppropriatenessInformativenessOursevaluation_protocol=LL...evaluation_protocol=LLM-as-a-judge2026.053.174.644.483.01T5-Zhevaluation_protocol=LL...evaluation_protocol=LLM-as-a-judge2026.052.654.144.012.3TPDialevaluation_protocol=LL...evaluation_protocol=LLM-as-a-judge2026.052.574.334.552.76TRIPDialevaluation_protocol=LL...evaluation_protocol=LLM-as-a-judge2026.052.4344.192.37