AI
AI 资讯
alishangtian.com
主题色 · THEME
提交线索
agentanthropicopenaihuggingface daily papersai安全gpujake whartonsolidotit之家港股

搜索：Tuning

共命中 50 条（服务端检索）
一篇读懂 PEFT 全家桶：LoRA、QLoRA、Prefix Tuning 与 Adapter 的选择
从全量微调的显存账本讲起：LoRA 的秩一分解数学与参数量核算、QLoRA 的 NF4 三件套、Prefix Tuning 与 Adapter 的取舍，附横向对比表与选型决策树。
大模型 精选 · 原创 · 今天 阅读 2·访客 2
Drift-Constrained Optimization: Only Direction Matters in Fine-Tuning Instruct Models
Fine-tuning instruct models often improves target performance while inducing behavioral drift from the reference model, …
行业动态 HuggingFace Daily Papers · 9-12 阅读 34·访客 33
CopilotKit Releases CLiF: A Continual Learning Platform That Turns Agent Traces and User Actions into Skills, Evals, and Fine-Tuning Data
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体 MarkTechPost · 10-15 阅读 74·访客 73
NVIDIA Announces DGX Spark 64GB: A 1-PetaFLOP Grace Blackwell Desktop for Local AI Agents, Fine-Tuning, and Inference
NVIDIA announced a new 64GB configuration of DGX Spark — from Acer, ASUS, Dell, Gigabyte, HP and MSI — its GB10-powered …
智能体 MarkTechPost · 10-3 阅读 110·访客 109
Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models
Striking a balance between helpfulness and safety remains a fundamental challenge in aligning large language models. To …
行业动态 HuggingFace Daily Papers · 9-4 阅读 18·访客 17
Follow the Gradient: Tracing Safety Degradation Back to Training Data
Fine-tuning Large Language Models (LLMs) on instruction data can quietly erode their safety, even when no training examp…
大模型 HuggingFace Daily Papers · 10-27 阅读 9·访客 9
SEAM: Spectral Embedding Alignment and Merging for Low-Resource TTS
Expanding the vocabulary of conditional text-to-speech (TTS) models typically requires extensive fine-tuning with paired…
智能体 HuggingFace Daily Papers · 10-19 阅读 45·访客 45
ImIR: Image-Instruction Tuning for All-in-One Image Restoration
Degradations vary widely across images, so a practical restoration system has to handle many degradation types with one …
行业动态 HuggingFace Daily Papers · 9-21 阅读 35·访客 35
Don't Mask the Environment: Observation Supervision Changes How Agents Explore Under RL
Agent trajectories record what an agent does and what happens next. Yet standard supervised fine-tuning (SFT) applies lo…
智能体 HuggingFace Daily Papers · 9-17 阅读 84·访客 82
CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents
Current Mixture-of-Agents (MoA) paradigms generally treat query routing and agent fine-tuning as separate processes, lim…
智能体 HuggingFace Daily Papers · 9-16 阅读 77·访客 76
联邦学习遇上大模型：分布式训练的隐私前沿
从 FedAvg 到差分隐私大模型训练，梳理联邦学习与大模型结合的技术现状：记忆化攻击、梯度泄露、DP-SGD 与联邦微调的进展与开放问题。
研究前沿 精选 · 原创 · 今天 阅读 4·访客 4
Prompt 的科学研究：提示词对模型行为的影响机制
提示词不是玄学而是可测量变量：FormatSpread 的 76 点波动、指令模板不稳定性、贝叶斯机制解释、软提示与自动提示优化、注入攻击与工程规范。
研究前沿 精选 · 原创 · 今天 阅读 0·访客 0
长尾知识与模型记忆：稀有知识的存储与检索机制
语言模型能存多少知识、稀有知识为何丢、检索头与 kNN-LM 等混合记忆机制如何补长尾：知识容量定律、PopQA 实证与参数外置路线全景。
研究前沿 精选 · 原创 · 今天 阅读 0·访客 0
知道自己不知道：大模型校准与不确定性量化前沿
从温度缩放到 GPT-4 的校准悖论，再到语义熵幻觉检测与弃答训练，梳理大模型「知道自己不知道」这条研究线的经典方法、最新进展与工程落地。
研究前沿 精选 · 原创 · 今天 阅读 0·访客 0
Deep Principle Releases Hydrogen-1.0: A 753B Open-Weight Model for Scientific Discovery
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态 MarkTechPost · 6天前 阅读 17·访客 17
从 SFT 到宪法 AI：对齐的技术路径图
大模型的「听话」不是天生的，是训练出来的：监督微调（SFT）教它格式，人类反馈强化学习（RLHF）教它偏好，宪法 AI（CAI）试着教它原则。本文讲清这条对齐技术路径的演进逻辑、每一代方法解决了什么又留下了什么，以及为什么 Anthropic 会说「AI 用 AI 的反馈来对齐 AI」。
研究前沿 精选 · 原创 · 6天前 阅读 46·访客 46
一篇读懂 SFT：指令微调怎么把基座模型变成助手
SFT（监督微调）是基座模型变成对话助手的第一步，也是最被误解的一步——它常被当成「教知识」，实际是「教格式与行为」。本篇拆解 SFT 的数据形态、训练细节（epoch、学习率、灾难性遗忘的边界）、常见的失败模式，以及「SFT 教的到底是不是能力」这个核心争议。
大模型 精选 · 原创 · 6天前 阅读 31·访客 30
微调、RAG 还是提示词：一张决策表选对适配方式
「我的场景该微调、上 RAG，还是写好提示词」——这个问题 80% 的答案可以在一张决策表里找到。三种适配方式改的是模型的三样东西：行为、知识、与当次指令。本篇用「你想改什么」做主轴拆解决策逻辑，给出各方式的成本与风险、混合架构的常见形态，以及一个五分钟的自测流程。
大模型 精选 · 原创 · 6天前 阅读 21·访客 21
操作系统内功 · 第 14 篇｜一次网络收包的内核之旅：从网卡 DMA 到 epoll 返回
P99 飙升、CPU 总量不高、却有一个核的软中断逼近 100%——问题几乎总出在收包路径。本文沿「网卡 DMA → 硬中断 → NAPI 轮询 → 软中断预算 → GRO → 协议栈 → socket 队列 → epoll 唤醒」的完整数据通路走一遍，讲清每一站的成本、丢包计数器与调优参数，最后给出可照抄的排查清单与分层定位表。
后端技术 精选 · 原创 · 11-2 阅读 71·访客 69
AiSOC（beenuar/AiSOC）：把本地大模型请进安全运营中心，每个结论都带可审计的账本
Cyble 资助的开源 AI SOC：OCSF 规范化、2530 条实测可执行检测规则、Ollama 本地模型分诊告警，提示词与工具调用全部落账本，响应默认必经人工审批。
开源项目 精选 · 用户投稿 · 11-2 阅读 60·访客 57
Order-Only On-Policy Distillation
On-policy distillation (OPD) trains a student on its own generations by matching the teacher's token probabilities. Thes…
行业动态 HuggingFace Daily Papers · 11-1 阅读 8·访客 8
NorthCUA Technical Report: Turbocharging Computer-Use Agents through Experience at Scale
Scaling computer-use agents from individual actions to complex software work requires experience that connects decisions…
智能体 HuggingFace Daily Papers · 11-1 阅读 11·访客 11
BYOD: Build Your Own DLM - Efficient Autoregressive-to-Diffusion Language Model Conversion
Unlike autoregressive (AR) models, diffusion language models (DLMs) permit parallel token updates, flexible generation, …
行业动态 HuggingFace Daily Papers · 10-30 阅读 8·访客 8
Where Do Vision Language Models Fail at 3D Reasoning?
Vision-language models can get individual judgments right while struggling to recover and consistently use the exact pro…
研究前沿 HuggingFace Daily Papers · 10-29 阅读 13·访客 13
帮孩子给情绪起名字：情绪粒度与语言的力量
孩子崩溃大哭时，一句「你是不是很委屈」不只是安慰，而是给大脑装了一个降噪器。本文从情绪粒度与「情绪标注」的神经证据讲起，梳理儿童情绪词汇的发展时间线，给出情感教练五步法与家庭词汇工具箱，并澄清「命名情绪会强化哭闹」等常见误区。
一叶一世界 精选 · 原创 · 10-29 阅读 118·访客 118
Thin, Light and Built for Everything: NVIDIA RTX Spark Windows Laptops Gain New Models, Agent Tools and Games
NVIDIA RTX Spark powers thin and light laptops built for everything that developers, creators and gamers do. RTX Spark l…
智能体 NVIDIA Blog · 10-28 阅读 14·访客 14
Anthropic co-founder Jack Clark moves his prediction for self-improving AI forward to 2027
Manuel Uth Oct 25, 2026 Nano Banana Pro prompted by THE DECODER A new benchmark called TasteVal measures how well AI sys…
研究前沿 The Decoder · 10-25 阅读 42·访客 42
NVIDIA Releases Nemotron 3.5 Bolt E4B 4GB: An Open Small Language Model for Local Voice Agents, Game Characters and Robots
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体 MarkTechPost · 10-24 阅读 69·访客 67
一篇读懂优化器：从 SGD、AdamW 到 Muon
十年来 AdamW 是大模型训练不容置疑的默认优化器，直到 Kimi K2 用它跑起万亿参数训练。本文从梯度下降的几何直觉讲起，拆解 AdamW 为什么统治了十年、Muon 的正交化更新为什么只需约 52% 的训练算力就能追平 AdamW，以及这轮换血里真正的工程细节与边界。
大模型 精选 · 原创 · 10-24 阅读 64·访客 63
NVIDIA Transforms European Retail, From Digital Discovery to Physical Operations
European retailers are using open models and physical intelligence across use cases — from a commerce platform cutting i…
大模型 NVIDIA Blog · 10-22 阅读 32·访客 32
NVIDIA Cosmos Opens New Tools to Train, Evaluate and Deploy World Models
To help developers rapidly build world models, which help physical AI systems understand and simulate the real world, NV…
行业动态 NVIDIA Blog · 10-22 阅读 23·访客 23
AnyGameRenderer: Entity-Consistent Generative Rendering Across Games and an Entity-Level Benchmark
AI coding agents can now build small playable games in minutes, but the results usually look like programmer art: gray b…
智能体 HuggingFace Daily Papers · 10-22 阅读 49·访客 49
Skeleton-Grounded Latent Action Model for Robot World Models
Robot world models simulate how environments evolve in response to robot actions. Latent action models enable learning d…
智能体 HuggingFace Daily Papers · 10-22 阅读 58·访客 57
Inherited, Not Invented: Tracing the Provenance of Wait in Reasoning Models
Reinforcement learning (RL) is credited with teaching reasoning models to write {wait} before reconsidering a step, yet …
研究前沿 HuggingFace Daily Papers · 10-22 阅读 21·访客 21
Bringing the Spark to Europe: Local Intelligence Powers Next Wave of Agents
At GTC Berlin, NVIDIA is showcasing its full lineup of local AI hardware. RTX Spark, which launched last week, powers pe…
智能体 NVIDIA Blog · 10-21 阅读 59·访客 59
Europe’s Industrial Leaders Reinvent Engineering and Manufacturing With NVIDIA
To increase productivity, accelerate innovation and create new opportunities for economic growth, Europe’s industrial le…
行业动态 NVIDIA Blog · 10-21 阅读 21·访客 21
Foundation4D: Towards a unified perception and reconstruction model for autonomous vehicles
We introduce Foundation4D, a unified feed-forward model for geometry estimation, 3D detection, semantic occupancy, and G…
行业动态 HuggingFace Daily Papers · 10-21 阅读 20·访客 20
Google Cloud and NVIDIA Announce General Availability of NVIDIA Vera Rubin Infrastructure
General availability of the instances is coming next month, with planned expansion of availability to Europe — bringing …
行业动态 NVIDIA Blog · 10-21 阅读 21·访客 21
LaLa-World: Scaling Latent World Models with Natural Language Actions
This paper shows that JEPA-style latent world models (LWMs) can learn across diverse domains and data sources by adoptin…
研究前沿 HuggingFace Daily Papers · 10-21 阅读 16·访客 16
NVIDIA Cloud Partners Scale Across Europe — Sovereign, Open and Built for Safe AI Agents
Demand from AI-natives is pulling the European AI cloud ecosystem forward faster than ever. Developers need compute clos…
智能体 NVIDIA Blog · 10-21 阅读 61·访客 61
MambaAct: Efficient Mamba Action Experts for Flow-Matching Vision-Language-Action Models
Vision-language-action (VLA) models have shown strong promise for generalist robotic manipulation, but their inference c…
智能体 HuggingFace Daily Papers · 10-16 阅读 43·访客 43
Nvidia's Nemotron 3.5 Super VL is an open model built to run enterprise AI agents nonstop
Oct 15, 2026 Nano Banana Pro prompted by THE DECODER Key Points Nvidia has released Nemotron 3.5 Super VL, an open multi…
智能体 The Decoder · 10-16 阅读 93·访客 92
Fish Audio Releases Drama 3, a 35B Open-Weight MoE Text-to-Speech Model
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
行业动态 MarkTechPost · 10-15 阅读 51·访客 51
Liquid AI Releases LFM2.5-24B-A2B: A 2B-Active MoE for On-Device Agents
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
智能体 MarkTechPost · 10-14 阅读 97·访客 97
共享符号而非向量：跨语言模型无需训练的事实通信方法 SYNC
提出 SYNC，一种基于共享词表索引的免训练接口，使异构语言模型无需微调或学习型翻译器即可进行多智能体间的事实通信。
研究前沿 HuggingFace Daily Papers · 10-14 阅读 29·访客 29
Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots
Oct 12, 2026 Nano Banana Pro prompted by THE DECODER Key Points A single language model improves itself through "Multi-A…
智能体 The Decoder · 10-13 阅读 105·访客 104
Sakana AI’s LLM Peer Review System Catches 73% of Core-Claim Errors
!(https://www.gstatic.com/images/branding/googleg/1x/googleg_standard_color_128dp.png)Add as a preferredsource on Google…
研究前沿 MarkTechPost · 10-11 阅读 425·访客 425
深度学习的数学基础：我们知道它能用，却说不清为什么
2017 年 Ali Rahimi 领奖时说「机器学习变成了炼金术」，LeCun 当场反驳；八年过去，模型能力涨了几个数量级，「为什么泛化」的数学解释仍然是一堆可解模型里的定理。本文盘点深度学习数学基础的现状：随机标签实验、双下降、NTK、grokking、scaling laws 各自被解释到了哪一步，哪一步仍然卡住，以及 2026 年「学习力学」的学科宣言算不算转折点。
研究前沿 精选 · 原创 · 10-10 阅读 509·访客 507
We’re putting too much faith in AI’s ability to say no
Ever since people first seriously contemplated giving machines an intelligence modeled on our own, there has never been …
行业动态 MIT Technology Review · 10-9 阅读 33·访客 33
Continually Post-training Small Models via Synthetic Self-distillation
Continual post-training faces a dual challenge of acquisition and retention: models must learn new capabilities while pr…
行业动态 HuggingFace Daily Papers · 10-9 阅读 13·访客 13