WeightLoop框架让GPT-6 Astra用实验反馈直接改写模型权重
WeightLoop: Weight-Space Programming with Experimental Feedback
让模型自己看着实验数据改自己的权重，不用重训练。用 GPT-6 Astra 做的，回放里 8 组测试赢了 7 组，思路挺新。
arXiv 论文介绍 WeightLoop 框架，属于权重空间编程（WSP）与递归自我改进（RSI）的结合。GPT-6 Astra 读取任务规范和模型实现后，生成一段有边界的可复用参数编辑程序。新测量数据到来时，程序找出预测偏差，用观测结果验证候选改动，再写回原模型。编辑器固定不变、预测器持续演化，更新过程不需要额外 LLM 调用或梯度重训练。离线回放中，定向编辑在 8 个端点预算设置中的 7 个提升了平均排序效用，对比随机编辑则是 8 个全部提升，并在 3 个序列任务上提高了平均最终发现数。
WeightLoop: Weight-Space Programming with Experimental Feedback
Weight-space programming (WSP) shows that language models can reason about and rewrite neural network weights; recursive self-improvement (RSI) motivates feedback-driven improvement. Models guiding experiments receive concrete feedback: completed assays reveal prediction errors and inform what to test next. We introduce WeightLoop, a framework that turns experimental feedback into persistent updates to deployed predictors. GPT-6 Astra inspects a task specification and the model implementation, then writes a bounded, reusable parameter-editing program. As new measurements arrive, the program identifies prediction discrepancies, validates candidate edits using observed outcomes, and writes the selected changes back to the original model. The editor remains fixed while the predictor evolves, enabling repeated updates without further LLM calls or gradient retraining while preserving the original inference interface. In offline replay, targeted edits improve mean ranking utility in seven of eight endpoint-budget settings (all eight versus random edits) and mean final discovery across three sequential tasks. WeightLoop connects WSP with an RSI-inspired improvement loop, making experimental evidence a direct driver of model change and subsequent decisions.