JETHICS provides 78,000 Japanese moral judgment examples, and evaluation shows GPT-4o averages about 0.71 accuracy while the best non-proprietary Japanese LLM reaches about 0.50.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
JETHICS: Japanese Ethics Understanding Evaluation Dataset
JETHICS provides 78,000 Japanese moral judgment examples, and evaluation shows GPT-4o averages about 0.71 accuracy while the best non-proprietary Japanese LLM reaches about 0.50.