human-evaluation
Find informative examples to efficiently (human)-evaluate NLG models.
Concept-Guided Chain-of-Thought (CGCoT) pairwise annotation using Large Language Models
