ResearchBenchmarksCuriosity-driven dialogue evaluation on Gemini-based dialoguesFollow3.21L1 ScoreCURIOBOT2.4822.6712.863.049Jun 21, 2026Evaluation ResultsMethodMethodLinksL1 ScoreL2 ScoreL3 ScoreL4 ScoreT1 ScoreT2 ScoreT3 ScoreT4 ScoreT5 ScoreT6 ScoreCURIOBOTModel=GeminiModel=Gemini2026.063.212.863.363.43.662.83.253.594.253.18BaselineModel=GeminiModel=Gemini2026.062.511.942.562.555.32.794.423.884.073.13