SAGE: Hierarchical LLM-Based Literary Evaluation through Ontology-Grounded Interpretive Dimensions

Tianyu Wang, Nianjun Zhou

arXiv:2605.07102 · 2026-05-11 공개 · arXiv · PDF

llm-evaluation ontology-grounded hierarchical-evaluation genre-hierarchy inter-rater-agreement text-generation-evaluation cultural-representation philosophical-depth

Abstract

Evaluating literary quality requires assessing interpretive dimensions such as cultural representation, emotional depth, and philosophical sophistication that resist straightforward computational measurement. We introduce SAGE, a hierarchical evaluation framework that decomposes literary quality into ontology-grounded interpretive dimensions assessed through structured large language model evaluation with multi-round iterative reflection and independent validation. We validate the framework on 100 short stories (50 canonical works, 30 pulp fiction, 20 LLM-generated narratives) across three analytical layers (cultural, emotional-psychological, existential-philosophical) using dual-mode assessment. Across 600 evaluations, the framework achieves 98.8% score convergence and greater than 94% inter-rater agreement, with near-perfect mode invariance between content-based and metadata-based evaluation. Statistical analysis reveals a consistent genre hierarchy (Canonical > Pulp > LLM, all p<0.001) with layer-specific discrimination: cultural critique and philosophical depth exhibit very large effect sizes (Cohen's d>2.4), while emotional representation shows smaller gaps (d=1.68), suggesting that affective patterns are more learnable from training data than critical stance or philosophical depth. Cross-layer correlations (r=0.649-0.683) confirm the three dimensions capture empirically distinguishable quality facets. These findings demonstrate that theory-driven LLM evaluation can achieve measurement-grade reliability and support systematic identification of where current generative models fall short of human literary production, with direct implications for scalable automated evaluation of open-ended text generation.

한국어 요약

📋 한 줄 요약

**[LLM 평가/NLP]** 문학 작품의 질적 평가를 온톨로지 기반 해석적 차원으로 분해해 LLM이 측정 수준의 신뢰도로 평가하도록 만든 계층적 프레임워크 SAGE를 제안.

🎯 핵심 기여도

💡 핵심 아이디어

"문학의 질"처럼 수치화가 어려운 해석적 개념도, 평가를 충분히 잘 정의된 하위 차원들로 계층화하고 LLM에게 구조화된 채점 절차를 강제하면, 인간 평가자 수준의 신뢰도로 측정 가능하다는 것이 핵심 가설이다.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: 개방형 텍스트 생성에 대한 자동 평가의 새로운 기준선을 제시하며, LLM이 어디서 인간 문학 생산에 미달하는지를 체계적으로 식별 가능하게 한다. **한계**: 100편이라는 표본 규모와 영어권 단편 위주 평가가 일반화 가능성을 제약하고, 평가자로 사용된 LLM 자체의 편향이 일부 결과에 잔존할 수 있다.

🚀 실용적 활용