ResearchBenchmarksDialogue Generation on DuRecDial ID 2.0 (test)Dialogue Generation on DuRecDial ID 2.0 (test)Follow2.9ProactivityOurs2.28642.44572.6052.7643May 12, 2026Evaluation resultsMethodMethodLinksProactivityCoherenceAppropriatenessInformativenessOursEvaluation Protocol=LL...Evaluation Protocol=LLM-as-a-judge2026.052.94.644.532.75T5-FlanEvaluation Protocol=LL...Evaluation Protocol=LLM-as-a-judge2026.052.54.354.082.2TRIPDialEvaluation Protocol=LL...Evaluation Protocol=LLM-as-a-judge2026.052.334.213.982.25TPDialEvaluation Protocol=LL...Evaluation Protocol=LLM-as-a-judge2026.052.314.1542.13