TaeHyun LEE

Beyond Correctness: Towards an Ethico Onto Epistemological Framework for the Evaluation of Large Language Models

The present generation of large language models is primarily evaluated through measures of factual accuracy, logical coherence, fluency, task completion and human preference. These criteria have undoubtedly accelerated progress in artificial intelligence. Yet they all presuppose...

Beyond Correctness: Towards an Ethico Onto Epistemological Framework for the Evaluation of Large Language Models

The present generation of large language models is primarily evaluated through measures of factual accuracy, logical coherence, fluency, task completion and human preference. These criteria have undoubtedly accelerated progress in artificial intelligence. Yet they all presuppose...