Abstract: In this article, we present BenchING, a new benchmark for evaluating large language models (LLMs) on their ability to follow structured output format instructions in text-based procedural ...

Many authors have long been engaged in the automatic generation of test items or stimuli, recognizing their potential for improving test efficiency, scalability, and psychometric quality. Pioneering ...

Abstract: Many current image restoration approaches utilize neural networks to acquire robust image-level priors from extensive datasets, aiming to reconstruct missing details. Nevertheless, these ...

Text-based games, management RPGs, and visual novels – of whatever kind – rarely make the most exciting previews. It's just not easy to build a rich interest in the world and characters of a story you ...