Politeness-controlled Text Generation on WIKIPOL benchmark leaderboard · SOTA2 Research

Politeness-controlled Text Generation on WIKIPOL

79.2AccuracySWAI

Evaluation Results

Method	Links
SWAI2026.01		79.2	70.9	84.96	66.37	57.7	58.2	61.6
SWAI2026.01		77.2	73	82.89	69.28	60.3	59.4	62.6
Baseline2026.01		56.7	56.9	56.79	57.1	74.8	35.1	35.1
Baseline2026.01		49.48	49.4	49.34	49.59	73.4	24.1	24.1