A Better LLM Evaluator for Text Generation: The Impact of Prompt Output Sequencing and Optimization
This work addresses prompt design issues for researchers and practitioners using LLMs to evaluate text generation, though it appears incremental.
The researchers tackled the challenge of creating effective prompts for LLM-based evaluation of open-ended text generation by experimenting with different prompt structures, finding that the order of presenting reasons and scores significantly influences LLM scoring.
This research investigates prompt designs of evaluating generated texts using large language models (LLMs). While LLMs are increasingly used for scoring various inputs, creating effective prompts for open-ended text evaluation remains challenging due to model sensitivity and subjectivity in evaluation of text generation. Our study experimented with different prompt structures, altering the sequence of output instructions and including explanatory reasons. We found that the order of presenting reasons and scores significantly influences LLMs' scoring, with a different level of rule understanding in the prompt. An additional optimization may enhance scoring alignment if sufficient data is available. This insight is crucial for improving the accuracy and consistency of LLM-based evaluations.