https://aibussin.com/tags/scm/ Recent content in SCM on AIBussin — AI applications, systems and books Hugo en-US Wed, 22 Oct 2025 20:30:36 +0100 https://aibussin.com/post/gap/ Wed, 22 Oct 2025 20:30:36 +0100 https://aibussin.com/post/gap/ <h2 id="-summary">🌌 Summary</h2> <p>What if the most valuable insights in AI evaluation aren&rsquo;t in model agreements, but in <strong>systematic disagreements</strong>?</p> <p>This post reveals that the &ldquo;gap&rdquo; between large and small reasoning models contains <strong>structured, measurable intelligence</strong> about how different architectures reason. We demonstrate how to transform model disagreements from a problem into a solution, using the space between models to make tiny networks behave more like their heavyweight counterparts.</p> <p>We start by assembling a high-quality corpus (10k–50k conversation turns), score it with a local LLM to create targets, and train both HRM and Tiny models under identical conditions. Then we run fresh documents through both models, collecting not just final scores but rich <strong>auxiliary signals</strong> (uncertainty, consistency, OOD detection, etc.) and visualize what these signals reveal.</p>