MegaTempQA is the largest temporal question answering dataset to date, containing over 250 million carefully curated question-answer pairs designed to evaluate temporal reasoning capabilities in large language models.
🎯 Key Features
🔢 Scale: 250+ million high-quality temporal QA pairs
📅 Coverage: Spans multiple centuries of historical events
🌍 Diversity: Covers events, entities, and timelines across various domains
🧠 Complexity: 16 distinct question types for comprehensive evaluation
📈 Multi-granular: Supports temporal reasoning at different time scales
🔗 Multi-hop: Enables complex inferential reasoning tasks
