SemScore: Automated Evaluation of Instruction-Tuned LLMs based on Semantic Textual Similarity

Instruction-tuned Large Language Models (LLMs) have recently showcased remarkable advancements in their ability to generate fitting responses to natural language instructions. However, many current works rely on manual evaluation to judge the quality of generated responses. Since such manual evaluation is time-consuming, it does not easily scale to the evaluation of multiple models and model variants. In this short paper, we propose a straightforward but remarkably effective evaluation metric called SemScore, in which we directly compare model outputs to gold …

automated cs.cl current evaluation generate generated instruction-tuned judge language language models large language large language models llms natural natural language quality responses scale semantic textual