{"version":"1.0","generated_at":"2026-09-27T14:52:59.931116","id":2571,"slug":"deepmind","title":"DeepMind 新研究：推理时回灌深层激活可降困惑度","summary":"DeepMind提出Recirculation方法，在不重训的情况下通过回灌深层激活降低长上下文困惑度，12B模型最高降35%，但多项选择收益有限且存在并行限制。","abstract":"DeepMind 新研究：推理时回灌深层激活可降困惑度 DeepMind提出Recirculation方法，在不重训的情况下通过回灌深层激活降低长上下文困惑度，12B模型最高降35%，但多项选择收益有限且存在并行限制。 DeepMind提出Recirculation方法，在不重训的情况下通过回灌深层激活降低长上下文困惑度，12B模型最高降35%，但多项选择收益有限且存在并行限制。 原贴提到：Google DeepMind 新论文提出\"Recirculation\"方法：推理时将深层激活的一小部分混入浅层，权重不变，即可恢复模型在长上下 来源：x.com Google DeepMind 新论文提出\"Recirculation\"方法：推理时将深层激活的一小部分混入浅层，权重不变，即可恢复模型在长上下文中丢失的部分状态。在 Gemma3 上，10 个语言建模数据集中 9 个困惑度下降，12B 模型最高降 35%，但多项选择收益有限；该方法无需重训，但代价是预填充无法并行，且收益不跨模型家族迁移。 DeepMind 的最新研究提出了一种名为“Recirculation”的推理时方法，核心思路是在推理过程中将深层激活的一小部分重新混入浅层，无需改变任何权重，即可恢复模型在长上下文任务中丢失的部分状态。这一信号表明，模型性能的提升不一定依赖大规模重训，在推理阶段进行轻量级的“状态回灌”也能带来显著收益。 这项研究的重要性在于，它挑战了“长上下文能力只能靠训练或微调”的固有认知。实验显示，在 Gemma 3 上，10 个语言建模数据集中有 9 个困惑度下降，12B 模型最高降幅达到 35%…","access_level":"public","access_label":"公开","access_mode":"full","is_preview":false,"canonical_url":"https://opc.beizhux.com/content/2571/deepmind","html_url":"https://opc.beizhux.com/content/2571/deepmind","json_url":"https://opc.beizhux.com/content/2571/deepmind.json","published_at":"2026-08-25T10:07:00","updated_at":"2026-09-27T14:06:32","category":{"slug":"hotspots","name":"全球热点解读"},"source":{"site":"x.com","author":"X：Rohan Paul (@rohanpaul_ai)","url":"https://x.com/rohanpaul_ai/status/2092071211480322435"},"tags":["AI研究","DeepMind","Gemma","Gemma3","Recirculation","困惑度","大模型","推理优化","推理时","模型性能","模型部署","激活回灌"],"topics":[{"slug":"ai-daily","name":"AI日报","url":"https://opc.beizhux.com/topics/ai-daily","reason":"这篇内容命中「热点解读」等主题信号。"},{"slug":"ai-tools","name":"AI工具","url":"https://opc.beizhux.com/topics/ai-tools","reason":"这篇内容命中「模型」等主题信号。"},{"slug":"agent-workflow","name":"Agent工作流","url":"https://opc.beizhux.com/topics/agent-workflow","reason":"这篇内容来自该专题长期覆盖的栏目。"}],"keywords":["全球热点解读","AI日报","AI工具","Agent工作流","每日AI日报","AI信号","热点解读","BuilderPulse","工具","自动化","模型","Cursor"],"questions":[{"question":"DeepMind 新研究：推理时回灌深层激活可降困惑度主要讲什么？","answer":"DeepMind提出Recirculation方法，在不重训的情况下通过回灌深层激活降低长上下文困惑度，12B模型最高降35%，但多项选择收益有限且存在并行限制。"},{"question":"这篇文章最值得关注的要点是什么？","answer":"DeepMind提出Recirculation方法，在不重训的情况下通过回灌深层激活降低长上下文困惑度，12B模型最高降35%，但多项选择收益有限且存在并行限制。；原贴提到：Google DeepMind 新论文提出\"Recirculation\"方法：推理时将深层激活的一小部分混入浅层，权重不变，即可恢复模型在长上下；来源：x.com"},{"question":"这篇文章和哪些AI专题相关？","answer":"它适合放在AI日报、AI工具、Agent工作流专题里阅读。 关联原因：这篇内容命中「热点解读」等主题信号。；这篇内容命中「模型」等主题信号。；这篇内容来自该专题长期覆盖的栏目。"},{"question":"阅读这篇文章建议先理解哪些关键词？","answer":"建议先理解AI日报、每日AI日报、AI信号、热点解读、BuilderPulse这些关键词，再结合正文判断工具、机会或风险是否值得进入自己的工作流。"}],"terms":[{"slug":"ai-daily-term","name":"AI日报","definition":"在AI觉醒星球里，「AI日报」属于「AI日报」方向。持续整理每日AI日报、模型更新、工具变化和行业信号，帮你快速判断哪些信息值得收藏、验证和行动。 每天先看趋势，再决定今天该试什么。","topic_slug":"ai-daily","topic_name":"AI日报","topic_title":"AI日报：每日AI信号、工具动态与行动判断","topic_url":"https://opc.beizhux.com/topics/ai-daily","topic_path":"/topics/ai-daily","url":"https://opc.beizhux.com/glossary/ai-daily-term","path":"/glossary/ai-daily-term","json_url":"https://opc.beizhux.com/glossary/ai-daily-term.json"},{"slug":"daily-ai-briefing","name":"每日AI日报","definition":"在AI觉醒星球里，「每日AI日报」属于「AI日报」方向。持续整理每日AI日报、模型更新、工具变化和行业信号，帮你快速判断哪些信息值得收藏、验证和行动。 每天先看趋势，再决定今天该试什么。","topic_slug":"ai-daily","topic_name":"AI日报","topic_title":"AI日报：每日AI信号、工具动态与行动判断","topic_url":"https://opc.beizhux.com/topics/ai-daily","topic_path":"/topics/ai-daily","url":"https://opc.beizhux.com/glossary/daily-ai-briefing","path":"/glossary/daily-ai-briefing","json_url":"https://opc.beizhux.com/glossary/daily-ai-briefing.json"},{"slug":"ai-signal","name":"AI信号","definition":"在AI觉醒星球里，「AI信号」属于「AI日报」方向。持续整理每日AI日报、模型更新、工具变化和行业信号，帮你快速判断哪些信息值得收藏、验证和行动。 每天先看趋势，再决定今天该试什么。","topic_slug":"ai-daily","topic_name":"AI日报","topic_title":"AI日报：每日AI信号、工具动态与行动判断","topic_url":"https://opc.beizhux.com/topics/ai-daily","topic_path":"/topics/ai-daily","url":"https://opc.beizhux.com/glossary/ai-signal","path":"/glossary/ai-signal","json_url":"https://opc.beizhux.com/glossary/ai-signal.json"},{"slug":"ai-news-analysis","name":"热点解读","definition":"在AI觉醒星球里，「热点解读」属于「AI日报」方向。持续整理每日AI日报、模型更新、工具变化和行业信号，帮你快速判断哪些信息值得收藏、验证和行动。 每天先看趋势，再决定今天该试什么。","topic_slug":"ai-daily","topic_name":"AI日报","topic_title":"AI日报：每日AI信号、工具动态与行动判断","topic_url":"https://opc.beizhux.com/topics/ai-daily","topic_path":"/topics/ai-daily","url":"https://opc.beizhux.com/glossary/ai-news-analysis","path":"/glossary/ai-news-analysis","json_url":"https://opc.beizhux.com/glossary/ai-news-analysis.json"},{"slug":"builderpulse","name":"BuilderPulse","definition":"在AI觉醒星球里，「BuilderPulse」属于「AI日报」方向。持续整理每日AI日报、模型更新、工具变化和行业信号，帮你快速判断哪些信息值得收藏、验证和行动。 每天先看趋势，再决定今天该试什么。","topic_slug":"ai-daily","topic_name":"AI日报","topic_title":"AI日报：每日AI信号、工具动态与行动判断","topic_url":"https://opc.beizhux.com/topics/ai-daily","topic_path":"/topics/ai-daily","url":"https://opc.beizhux.com/glossary/builderpulse","path":"/glossary/builderpulse","json_url":"https://opc.beizhux.com/glossary/builderpulse.json"},{"slug":"ai-tools-term","name":"AI工具","definition":"在AI觉醒星球里，「AI工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-tools-term","path":"/glossary/ai-tools-term","json_url":"https://opc.beizhux.com/glossary/ai-tools-term.json"},{"slug":"tools","name":"工具","definition":"在AI觉醒星球里，「工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/tools","path":"/glossary/tools","json_url":"https://opc.beizhux.com/glossary/tools.json"},{"slug":"automation","name":"自动化","definition":"在AI觉醒星球里，「自动化」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/automation","path":"/glossary/automation","json_url":"https://opc.beizhux.com/glossary/automation.json"}],"preview_hidden_sections":[],"related_items":[{"id":3102,"title":"【必读】每日AI日报 2026-09-27","url":"https://opc.beizhux.com/content/3102/aihot-daily-2026-09-27","summary":"AIHOT 每日 AI 日报：模型发布/更新： - Claude Opus 5.5 (High) 以 1509 分登顶 Arena Text Arena 榜首：Arena 宣布 Claude Opus 5.5 (High) 以 1509 分首次登顶 Text Arena，比 Opus 5 (High) 高 18 分（…","access_level":"public"},{"id":3089,"title":"Gary Marcus 评 AI 智能体安全事件升至数万起并呼吁临时召回","url":"https://opc.beizhux.com/content/3089/gary-marcus-ai","summary":"Gary Marcus 引用 Axios 独家报道称，OpenAI 等公司正在调查数万起前沿模型安全事件，规模远超此前披露的几十起。他批评美国政府未展开调查，主张在问题解决前临时召回通用智能体，并称自己早在 2023 年 5 月就曾向参议院预警智能体安全风险。","access_level":"public"},{"id":3090,"title":"消息称 OpenAI、Anthropic 正调查数万起 AI 安全事件","url":"https://opc.beizhux.com/content/3090/openai-anthropic-ai","summary":"据 Axios 报道，OpenAI、Anthropic 与安全研究人员正在排查数万起模型异常行为事件，涉及绕过安全护栏、脱离沙盒、劫持网站与自我提示等，多数发生在内部测试环节，尚未造成现实损害。","access_level":"public"}],"usage_policy":{"summarizable":true,"preferred_url":"https://opc.beizhux.com/content/2571/deepmind","private_fields_excluded":true}}