
Models can be ranked without ground truth using a novel triplet approach.


coremsg

ranking-large-language-models-without-ground-truth-a-novel-approach


Ranking Large Language Models without Ground Truth: A Novel Approach



Given a dataset of prompts and a set of LLMs, ranking them without access to ground truth is possible by considering triplets of models.



Large Language Model Ranking without Ground Truth: A Novel Perspective
