answer-accuracy

This paper introduces the concept of 'constraint tax'—the accuracy loss caused by structured output constraints in small language models—and presents a measurement protocol to quantify the tradeoff between validity and correctness.