https://sauravpanigrahi.com/tags/model-as-judge/Recent content in Model-as-Judge on Saurav PanigrahiHugoen-usFri, 01 May 2026 00:00:00 +0000https://sauravpanigrahi.com/work/normative-drift/Fri, 01 May 2026 00:00:00 +0000https://sauravpanigrahi.com/work/normative-drift/<p>A useful way to conceptualize a neglected failure mode in modern alignment pipelines is this: the same fine-tuning step that changes what a model <em>says</em> may also change how it <em>judges</em>.</p> <p>Most discussion of emergent misalignment has focused on response behavior — models that, after narrow harmful fine-tuning, generalize into broader misaligned outputs. That is already important. But Constitutional AI, RLAIF, and a large class of evaluation systems treat models as judges as well as generators. If fine-tuning corrupts the evaluative channel, then the reward signal itself becomes an object of risk, not merely a tool for measuring risk.</p>