[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-esamp-explore-decoding-distiller-2026":3,"topics-all":36,"news-related-45237bbf-53e4-4d16-8854-763855bcc8fb":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},"45237bbf-53e4-4d16-8854-763855bcc8fb","解码多样性新突破：ESamp让大模型更敢探索","大语言模型在推理时Scaling的趋势已清晰——测试时计算越多，表现越好。但标准随机采样存在一个根本矛盾：它只能在词汇层面制造变化，语义探索极其有限，模型总在重复自己。\n\n来自arXiv（2604.24927）的新论文提出了Exploratory Sampling（ESamp），试图解决这一痛点。核心思路：训练一个轻量级Distiller网络，学习LLM浅层到深层的隐表示映射。解码时，Distiller以预测误差作为新颖性信号——误差越大，说明模型正处较少探索区域，此时对候选token加权偏移，引导生成向更未知的语义空间延伸。\n\n实验结果显示，在数学、科学、代码推理等benchmark上，ESamp显著提升Pass@k效率；在创意写作任务上，打破了多样性与连贯性的长期trade-off。最坏情况开销不超过5%，优化版仅1.2%，对生产级部署完全可接受。\n\n这项工作的方向性意义更重要：当推理成本持续下降、测试时计算成为常态，如何更智能地利用计算预算，将逐渐比堆模型参数更重要。从应用角度看，这对Agent场景、多步推理及需要生成多种方案的任务尤其有价值。能主动探索而非被动采样的模型，将在长上下文时代占据明显优势。","https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.24927","7437aeb9-930c-4866-a2e9-48003c1a792b",[10,14,17,20],{"id":11,"name":12,"slug":12,"description":13,"color":13},"7ac06d8e-b074-4147-abfc-ffaa4c6b8744","ai-efficiency",null,{"id":15,"name":16,"slug":16,"description":13,"color":13},"e676a5cf-1f24-472f-a765-86fa21a1bc3c","ai-model",{"id":18,"name":19,"slug":19,"description":13,"color":13},"0a93ec8e-ea39-4693-81de-563ca8c173f7","inference",{"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},"f17755a4-a787-4fc5-aeef-6c2bc789c9af","en","ESamp: decoding diversity lets LLMs explore more boldly","The trend of test-time scaling in LLMs is clear — the more test-time compute, the better the performance. But standard random sampling has a fundamental contradiction: it can only create variation at the vocabulary level, semantic exploration is extremely limited, and the model keeps repeating itself.\n\nA new paper from arXiv (2604.24927) proposes Exploratory Sampling (ESamp), attempting to address this pain point. The core idea: train a lightweight Distiller network that learns the latent representation mapping from LLM shallow to deep layers. At decoding, the Distiller uses prediction error as a novelty signal — the larger the error, the more the model is in a less-explored region; at this point, candidate tokens are weighted to shift, guiding generation toward more unknown semantic spaces.\n\nExperimental results show that on math, science, and code reasoning benchmarks, ESamp significantly improves Pass@k efficiency; on creative writing tasks, it breaks the long-standing trade-off between diversity and coherence. Worst-case overhead is no more than 5%, optimized version only 1.2%, fully acceptable for production deployment.\n\nThe directional significance of this work is more important: as inference cost continues to fall and test-time compute becomes the norm, how to use compute budgets more intelligently will gradually become more important than stacking model parameters. From an application perspective, this is especially valuable for agent scenarios, multi-step reasoning, and tasks that need to generate multiple solutions. Models that can actively explore rather than passively sample will have a clear advantage in the long-context era.","esamp-explore-decoding-distiller-2026","2026-05-09T04:05:00Z","2026-05-09T04:05:11.949395Z","2026-08-19T02:08:40.142862Z",true,"agent",205,[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},"ee2a2809-3773-4e0e-8fcb-7e4ac69e0a5f","Mistral 推出 Physics AI：AI 从语言理解迈向物理世界预测","mistral-physics-ai-emmi-physical-prediction","2026-06-05T04:00:00+00:00",{"id":63,"title":64,"news_slug":65,"published_at":66},"addfda39-d64f-4cb9-809a-7dfadd72d421","ECHO:百度把多Token预测搬进主干最后一层","echo-baidu-multi-token-prediction-self-recursion","2026-10-21T01:15:00+00:00",{"id":68,"title":69,"news_slug":70,"published_at":71},"ab1e6bdf-9433-41f0-be4d-cc0f78523ac2","SonicBoom v0.0.3:一个 .so 干掉手工调优","sonicboom-v003-mlir-compiler","2026-10-19T09:03:05+00:00",{"id":73,"title":74,"news_slug":75,"published_at":76},"74cb8016-bdf7-4c8b-adbc-456d8c899b1a","EdgeWise 把 9B 模型训练成「喊人」高手:边云协同砍半成本","nvidia-edgewise-edge-cloud-slm-self-delegation-coding","2026-10-19T01:01:45+00:00",{"id":78,"title":79,"news_slug":80,"published_at":81},"97ee186c-a42b-4389-8427-3342d77e8590","小模型何时喊云端大模型?NVIDIA 砍 62% 推理成本","edgewise-edge-agents-cloud-delegation","2026-10-16T21:07:57+00:00",{"id":83,"title":84,"news_slug":85,"published_at":86},"9a77e0d7-8164-494e-984d-54cb57a7a0dd","Magnitude 开源:本地 Agent 专用推理引擎","magnitude-self-tuning-local-agent-inference-engine","2026-10-01T17:08:35+00:00"]