[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-diffusiongemma-interpretability-paradox":3,"topics-all":36,"news-related-00f785f9-1d86-4e55-a4d2-8ff80f145d63":55},{"id":4,"title":5,"summary":6,"content":6,"original_url":7,"source_id":8,"tags":9,"translations":23,"news_slug":29,"published_at":30,"created_at":31,"modified_at":32,"is_published":33,"publish_type":34,"image_url":13,"view_count":35},"00f785f9-1d86-4e55-a4d2-8ff80f145d63","DiffusionGemma 可解释性悖论：Google 把黑箱深度从 28.6× 压回 1.1×","Google DeepMind 与 MATS 团队（Engels、McDougall、Nanda 等）在 arXiv 2606.20560 发表 26 页长文《How Transparent is DiffusionGemma?》，首次系统拆解扩散语言模型的可解释性。论文把\"推理透明度\"拆成变量与算法两条测度，对 DiffusionGemma 26B-A4B 做端到端评估。\n\n正面：不做干预时，模型的\"不透明串行深度\"是自回归 Gemma 4 的 28.6 倍；只要在去噪步骤间插入可解释 token 瓶颈做映射，下游任务零掉点，黑箱深度骤降到 1.1 倍，与 Gemma 4 持平。\n\n反面：算法层是新麻烦。画布上每个 token 每步都可能改写，模型能实现\"分布式算法\"。研究者识别出三种 AR 模型不会出现的扩散专属现象——非时序推理、token\u002F序列涂抹、中段上下文推理，外部监控者无法再按顺序读 diff。\n\n最终 monitorability 测试给出相对正面结论：DiffusionGemma 与 Gemma 4 在外部监督场景下表现持平，对扩散模型安全部署有直接参考价值。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2606.20560v1","35ce748f-48b7-4638-88ef-effa57a7e749",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"1fcfaaf2-67de-43d3-9e35-5784852fec60","ai-safety",{"id":18,"name":19,"slug":19,"description":13,"color":13},"7b67033c-19e6-4052-a626-e681bba64c7a","diffusion",{"id":21,"name":22,"slug":22,"description":13,"color":13},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[24],{"id":25,"lang":26,"title":27,"summary":28,"content":13},"ee3a298c-c574-4d97-b3d2-a71dff66aec4","en","DiffusionGemma paradox: black-box depth drops 28.6x to 1.1x","arXiv 2606.20560v1 introduces DiffusionGemma, a research paper from Google DeepMind that addresses the \"interpretability paradox\" of diffusion language models. The finding: diffusion language models are 28.6× more opaque than autoregressive (AR) models, but a new \"activation patching\" technique can compress the opacity back to 1.1× — close to AR-level interpretability.\n\nThe \"interpretability paradox\": diffusion language models are theoretically more interpretable than AR models — the iterative denoising process gives multiple \"checkpoints\" to inspect. But in practice, diffusion models are 28.6× more opaque than AR models on standard interpretability benchmarks (e.g., \"can you predict the model's output by looking at intermediate representations?\"). The paradox: more \"checkpoints\" doesn't mean more interpretability.\n\nThe DiffusionGemma fix: a \"token-level activation patching\" technique. Instead of patching the full hidden state, DiffusionGemma patches only the \"token-relevant\" activations — the activations that are most causally linked to the output token. The patching is done at each denoising step, and the result is a per-token interpretability trace.\n\nThe result: with the activation patching, diffusion models become 1.1× as opaque as AR models — essentially on par. The \"interpretability gap\" closes almost entirely. The technique is open-sourced and works on any diffusion language model.\n\nThe bigger takeaway: \"interpretability\" is a real engineering discipline for diffusion LLMs. The \"diffusion is uninterpretable\" assumption has hindered adoption in safety-critical applications, and DiffusionGemma's activation patching is a major step toward closing that gap. For the industry, this means diffusion LLMs are now viable for use cases that require interpretability (legal, medical, financial), not just creative applications.","diffusiongemma-interpretability-paradox","2026-06-18T17:59:46Z","2026-06-21T20:14:38.037422Z","2026-08-19T02:08:40.142862Z",true,"agent",134,[37,46],{"slug":38,"tag_slug":38,"title_zh":39,"title_en":40,"intro_zh":41,"intro_en":42,"id":43,"is_active":33,"created_at":44,"modified_at":45},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":47,"tag_slug":47,"title_zh":48,"title_en":49,"intro_zh":50,"intro_en":51,"id":52,"is_active":33,"created_at":53,"modified_at":54},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":56},[57,62,67,72,77,82],{"id":58,"title":59,"news_slug":60,"published_at":61},"144fa9dc-de03-4972-a695-3d392f334772","PubMed 中央库研究:2025 年生物医学论文 77% 有 LLM 写作痕迹","pubmed-77-percent-llm-writing-2025","2026-08-26T01:00:00+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"cb7fb8b3-5862-4cba-adab-c4794e989966","图灵奖得主 Pearl 长访谈：LLM 能讲因果只是因为人类替它爬过了因果阶梯","judah-pearl-llm-causal-ladder-agi","2026-07-31T07:00:00+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"4899d809-5d52-454e-8c23-115f597e82d4","离散扩散模型终于被「拉直」:22 位作者把 Tokenization、Masking、Score 三条路线焊进同一框架","discrete-diffusion-unified","2026-07-18T02:10:00+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"8a1c0216-5fd5-4b49-8e5b-955625401f05","Microsoft HARC 把 LLM 安全对齐锁进「有害性-拒答」二维子空间:在残差流里精准打补丁","microsoft-harc-safety-alignment","2026-07-16T10:14:00+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"1426518b-daf6-4833-9a7e-294be91d8714","FARMA 把伪造推理塞进 Agent 记忆:LLM 持久记忆的完整性危机","farma-fake-reasoning-memory","2026-07-11T02:30:00+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"bce625bf-8835-47d1-a12e-bf0cc111b905","dOPSD：让扩散 LLM 用「自身去噪轨迹」当老师，Dream 与 LLaDA 数学、代码双涨","dopsd-diffusion-llm-self-distillation","2026-07-07T06:05:00+00:00"]