ResearchBenchmarksMulti-turn Dialogue Evaluation on MT-Bench Turn-2Multi-turn Dialogue Evaluation on MT-Bench Turn-2Follow1.9Writing ScoreLoRA1.3541.495751.63751.77925Feb 23, 2024Evaluation resultsMethodMethodLinksWriting ScoreRoleplay ScoreReasoning ScoreMath ScoreCoding ScoreExtraction ScoreStem ScoreHumanities ScoreAverage ScoreLoRABackbone=LLaMA-2, # Pa...Backbone=LLaMA-2, # Param=8.39M2024.021.95.82.11.62.551.2223.15.52.994FTBackbone=LLaMA-2, # Pa...Backbone=LLaMA-2, # Param=6739M2024.021.6675.9382.2221.722.1113.25.33.021REDBackbone=LLaMA-2, # Pa...Backbone=LLaMA-2, # Param=0.26M2024.021.3755.52.4441.4442.1251.7535.8752.946