AIML-TUDA/SLR-Bench-Spanish

SLR-Bench-Spanish 是 SLR-Bench 数据集的西班牙语版本，旨在评估和训练大型语言模型（LLMs）在西班牙语中的逻辑推理能力。该数据集包括自然语言任务提示、可执行的验证程序（用于自动评估）和潜在的地面真实规则。数据集分为 20 个复杂度级别，分为 4 个层级：基础、简单、中等和困难。数据集具有自动任务生成、可编程和可扩展的任务创建、符号化和自动化的评估功能，并支持课程学习。

SLR-Bench-Spanish is the Spanish version of the SLR-Bench dataset, designed for evaluating and training Large Language Models (LLMs) in logical reasoning in Spanish. It includes natural language task prompts, executable validation programs for automatic evaluation, and latent ground-truth rules. The dataset is structured into 20 complexity levels grouped into 4 tiers: basic, easy, medium, and hard. The dataset features automatic task generation, programmable and scalable task creation, symbolic and automated evaluation, and supports curriculum learning.