SurveyReview Benchmark for Human-Aligned LLM Survey Evaluation
August 11, 2026
SurveyReview provides a multi-dimensional benchmark consisting of 675 annotated survey papers and 1,630 peer-review reports. It quantifies LLM alignment with human reviewers across four metrics: Readability, Criticalness, Comprehensiveness, and Structure.
HOW THIS AFFECTS YOU
●
builderThis framework provides a structured way to evaluate automated content generation pipelines.
●
researcherYou can use this dataset to benchmark how well LLM-as-a-judge methods align with human academic standards.