NOLLI: A Difficulty-Calibrated Puzzle Benchmark for Diagnosing the English-Korean Performance Gap

2026-08-04

AI Summary

This study investigates the root causes of performance disparities between English and Korean tasks in large language models, focusing on the effects of writing systems, cultural context, and task complexity. To this end, we introduce NOLLI, a procedurally generated bilingual puzzle benchmark comprising 15 puzzle types (25 tasks, 7,500 instances), designed with behaviorally calibrated difficulty to ensure unique solutions and reproducibility. The benchmark features a three-tiered task structure: translation-matched pairs, Jamo-based adaptations, and culturally grounded Korean-only tasks. Innovatively replacing structural scale with behavioral calibration to define difficulty, our work introduces the first tasks explicitly leveraging Hangul’s Jamo components and Korean cultural knowledge. Experiments across 12 capable models reveal negligible performance gaps (±10%) on translation-matched tasks, but a pronounced 68.7% gap on tasks involving the Korean writing system. Notably, performance on Jamo composition tasks strongly predicts success on cryptographic reasoning, highlighting sub-syllabic multi-step inference as a critical bottleneck.

Abstract

We introduce NOLLI, a procedurally generated English-Korean puzzle benchmark designed to diagnose where Korean performance gaps arise. It comprises 15 puzzle types (25 tasks; 7,500 items), with every instance seed-regenerable, verified to have a unique solution, and scored deterministically. Rather than equating harder with bigger, we calibrate difficulty behaviorally, tuning each generator until a fixed reference model lands in target accuracy bands. Its three-level design combines matched direct translations, script adaptations over Hangul jamo (sub-syllabic letters), and Korean-only tasks grounded in Korean culture or orthography. We evaluate 15 frontier, open-weight, and Korean-developed models; among the 12 above a 3% overall-accuracy floor, matched English-Korean accuracy is statistically equivalent within a +/- 10 pp margin (TOST), suggesting little cost from presentation language alone. Writing-system-intensive tasks show sharper gaps: Korean Cipher falls behind English by up to 68.7 pp, whereas Cryptarithmetic over the same jamo shows no systematic penalty, and Jamo Composition accuracy predicts Korean Cipher accuracy. These contrasts are diagnostic rather than causal, consistent with difficulty in multi-step sub-syllabic execution. Korean-only tasks separate rule-application deficits, which vary in sign, from a Kinship deficit positive in all 12. Finally, a salient size measure fails to grow from Easy to Hard in 7 of 15 types, making structural size an unreliable proxy for empirical difficulty.

Problem

• performance gap
• English-Korean
• puzzle benchmark
• writing system
• difficulty calibration

Innovation

• procedurally generated benchmark
• difficulty calibration
• English-Korean performance gap
• jamo-based tasks
• behavioral difficulty tuning