MindGYM: What Matters in Question Synthesis for Thinking-Centric Fine-Tuning?
This work addresses the problem of enhancing reasoning capabilities in large AI models for researchers and practitioners, offering a scalable and self-challenging approach that reduces human intervention, though it is incremental as it builds on existing data synthesis paradigms.
The paper tackles the challenge of enabling large foundation models to acquire transferable, structured thinking abilities by proposing MindGYM, a thinking-centric data synthesis framework for question generation, which results in synthetic data with 16.7% higher average quality and 67.91% lower quality variance compared to baselines, and improves performance on reasoning benchmarks by up to 16% using only 400 data samples.
Large foundation models face challenges in acquiring transferable, structured thinking abilities, especially when supervised with rigid templates or crowd-annotated instruction datasets. Unlike prior approaches, we focus on a thinking-centric data synthesis paradigm that enables models to evolve through self-generated, cognitively guided data. We propose MindGYM, a structured and scalable framework for question synthesis, composed of: (1) Cognitive Thinking Process Injection, which infuses high-level reasoning objectives to shape the model's synthesis behavior; (2) Seed Single-Hop Question Synthesis, generating atomic questions from diverse semantic types to encourage broader thinking; and (3) Challenging Multi-Hop QA Synthesis, composing more complex multi-hop questions based on QA seeds for deeper reasoning. Detailed analysis shows that synthetic data generated by our method achieves 16.7% higher average quality and 67.91% lower quality variance compared to baseline sources, highlighting that both high-quality and self-contained data are essential for effective, thinking-oriented fine-tuning. MindGYM improves performance on six reasoning benchmarks, achieving gains of up to 16% on MathVision using only 400 data samples, and generalizable improvements across different model sizes and architectures. MindGYM underscores the viability of self-challenging mechanisms in refining large model capabilities while minimizing human intervention and resource demands. Code and data are released to promote data-centric research into self-evolving foundation models driven by their internal reasoning capabilities.