DateLogicQA: Benchmarking Temporal Biases in Large Language Models

Gagan Bhatia, MingZe Tang, Cristina Mahanta +1

This paper introduces DateLogicQA, a benchmark with 190 questions covering diverse date formats, temporal contexts, and reasoning types. We propose the Semantic Integrity Metric to…