评估大型语言模型在图表问答中的提示策略

Prompting strategies significantly affect the reasoning performance of large language models (LLMs), yet their role in chart-based question answering remains underexplored. A systematic evaluation of four widely used prompting paradigms (Zero-Shot, Few-Shot, Zero-Shot Chain-of-Thought, and Few-Shot Chain-of-Thought) is presented, conducted on the ChartQA dataset. The framework operates exclusively on structured chart data, isolating prompt structure as the only experimental variable, and evaluates performance using two metrics: Accuracy and Exact Match. Results from 1,200 diverse ChartQA samples indicate that Few-Shot Chain-of-Thought prompting consistently yields the highest accuracy (up to 78.2%), particularly on reasoning-intensive questions, while Few-Shot prompting improves format.