SAGE achieves 98.8% score convergence and over 94% inter-rater agreement on 100 stories, statistically confirming canonical works outperform pulp fiction and LLM-generated narratives especially in cultural critique and philosophical depth.
The prompt explicitly names non-Western traditions the evaluator must not overlook: ROUND 2/5: PROJECTION BIAS CHECK
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SAGE: Hierarchical LLM-Based Literary Evaluation through Ontology-Grounded Interpretive Dimensions
SAGE achieves 98.8% score convergence and over 94% inter-rater agreement on 100 stories, statistically confirming canonical works outperform pulp fiction and LLM-generated narratives especially in cultural critique and philosophical depth.