LookAlike: Consistent Distractor Generation in Math MCQs

Large language models (LLMs) are increasingly used to generate distractors for multiplechoice questions (MCQs), especially in domains like math education.However, existing approaches are limited in ensuring that the generated distractors are consistent with common student errors.We propose LOOKALIKE 3 , a method that improves error-distractor consistency via preference optimization.Our two main innovations are: (a) mining synthetic preference pairs from model inconsistencies, and (b) alternating supervised fine-tuning (SFT) with Direct Preference Optimization (DPO) to stabilize training.Unlike prior work that relies on heuristics or manually annotated preference data, LOOKALIKE uses its own generation inconsistencies as dispreferred samples, thus enabling scalable and stable training.Evaluated on a real-world dataset of 1,400+ math MCQs, LOOKALIKE achieves 51.6% accuracy in distractor generation and 57.2% in error generation under LLM-as-a-judge evaluation, outperforming an existing state-of-the-art method (45.6% / 47.7%).These improvements highlight the effectiveness of preference-based regularization and inconsistency mining for generating consistent math MCQ distractors at scale.7 of a group of students are boys.What would be a possible ratio of boys to girls?Key 3 : 4