编程模型过度编辑代码问题
This is a weird behavior in coding models and something worth looking into. It turns that some mode...
CROCODIL框架解决模型间代码编辑冲突，用双奖励机制减少过度修改。
研究发现不同模型会过度编辑其他模型生成的代码。CROCODIL框架通过相似性奖励和执行奖励的乘积机制减少这种行为。该研究测量了模型编辑其他模型代码时的变化情况。不同训练数据导致不同风格偏好，模型对"外来"代码的编辑往往过度。
This is a weird behavior in coding models and something worth looking into. It turns that some mode...
This is a weird behavior in coding models and something worth looking into. It turns that some models over-edit code that another models wrote. There is a high chance that your repo now has commits from more than one model, and that changes how each of them edits. Researchers measured what happens when one model edits code another model wrote. Different training data produces different stylistic preferences, and models make more edits, often excessive ones, on foreign code than on their own. CROCODIL is a post-training framework that reduces that behavior. A similarity reward penalizes large changes and an execution reward scores build and test success, and the two are multiplied rather than added. That product stops the policy from shrinking edits by simply failing the task. Paper: academy.dair.ai/papers/crocodi… 💬 12 🔄 7 ❤️ 25 👀 3098 📊 18 ⚡
