论文
71AI 评分(两次独立评分的平均)

斯坦福等机构论文:27B 开源模型可通过推理编辑自身权重自我改进

原标题:New Stanford+other top labs paper finds that a 27B open model can improve itself by editing its own weights through reasoning, with no tr...

Rohan Paul@rohanpaul_ai发布时间 (北京时间)

斯坦福等机构的新论文发现,27B 开源模型能在不训练、不用梯度的情况下,通过推理编辑自身权重实现自我改进。研究者给 Qwen3.5-27B 开放自身权重、带评分的练习题答案和一个 Python 会话并屏蔽梯度,模型对比答完与耗尽 token 两种答案的隐藏状态,削弱区分二者的方向。

中文译文尚未完成,先为你展示原文。

New Stanford+other top labs paper finds that a 27B open model can improve itself by editing its own weights through reasoning, with no training or gradients.

Improving a deployed model today means another training run on large clusters. Prompts and extra inference compute help, but the gain disappears once the task ends.

They gave Qwen3.5-27B its own weights, its graded practice answers, and a Python session, with gradients blocked. It compared its hidden states on answers that finished against answers that ran out of tokens, then weakened the direction separating them.

Its own edit raised accuracy from 57.2% to 63.8% on 490 unseen problems, in a test fixed in advance. The 2B model made no edit, and the 4B and 9B models broke themselves.

If you run a large open model, let it propose weight edits for its own failures, but keep only edits that pass held-out tests.

推理能力开源生态数据与训练论文研究#论文/研究#推理#数据/训练#开源生态