A Judge-free LLM Open-ended Generation Benchmark Based on the Distributional Hypothesis
This work addresses the problem of efficiently evaluating large language models for open-ended text generation, which is significant for natural language processing researchers and developers.
The authors tackled the challenge of evaluating open-ended text generation of large language models, achieving a strong correlation with human-based evaluations while requiring significantly fewer computational resources. Their benchmark uses 50 question and reference answer sets and achieves comparable results to GPT-4-based evaluations.
Evaluating the open-ended text generation of large language models (LLMs) is challenging because of the lack of a clear ground truth and the high cost of human or LLM-based assessments. We propose a novel benchmark that evaluates LLMs using n-gram statistics and rules, without relying on human judgement or LLM-as-a-judge approaches. Using 50 question and reference answer sets, we introduce three new metrics based on n-grams and rules: Fluency, Truthfulness, and Helpfulness. Our benchmark strongly correlates with GPT-4o-based evaluations while requiring significantly fewer computational resources, demonstrating its effectiveness as a scalable alternative for assessing LLMs' open-ended generation capabilities.