1
丹麦政府推出紧急方案引导学生合理使用 AI，遏制“AI 作弊”行为
IT之家 8 月 28 日消息，据 CCTV 国际时讯报道，在今年夏天的考试季，丹麦多所高中 AI 作弊案例明显增多。随着生成式人工智能（AI）进入校园，学生利用 AI 完成作业甚至在考试中作弊的问题，给传统教育和考试制度带来新挑战。 为此，丹麦政府日前推出紧急方案，通过增加口头答辩、监控学生考试电…
2
Move by Move: Measuring and Steering How LLMs Conduct Psychotherapy
拆解LLM心理治疗每步动作，精准测量并引导对话走向，让AI咨询更可控。
arXiv:2608.21325v1 Announce Type: new Abstract: Users increasingly turn to large language models for emotional support, yet little is known about how …
3
微软应用引导 Chrome / Firefox 等浏览器设 Bing 为默认搜索引擎，相关扩展拥有 500 万用户
微软靠一款500万用户的扩展悄悄引导Chrome和Firefox用户改用Bing，这招挺有看点
IT之家 8 月 22 日消息，科技媒体 Windows Latest 昨日（8 月 21 日）发布博文，报道称微软推出独立应用程序， 引导用户将 Bing 设置为 Edge、Firefox、Chrome 和 Brave 等浏览器默认搜索引擎。 该应用名为 Microsoft Recommended…
4
Win11 旧版 Media Player 再遭削弱：微软引导用户使用内存占用更高的新版应用
IT之家 8 月 19 日消息，据 Windows Latest 今日报道，微软正逐步削弱 Windows 经典媒体播放器 Windows Media Player 的功能。 Windows 观察人士 Xeno Panther 发现，在 Win11 最新实验版中，旧版播放器的字幕控制选项已被阉割，用…
5
Can Large Language Models Explain Flight Safety Events? A Prior-Guided Semantic LLM-based Approach
大模型如何解读飞行安全事件？这篇论文提出先验引导语义方法，为预测性安全分析开辟新路径。
arXiv:2608.18017v1 Announce Type: new Abstract: Improving flight safety with flight data requires not only accurate detection of risk events, but more…
6
TraVEL: Trajectory-Guided Video Embedding Learning for Driving-Video Retrieval
用车辆轨迹引导视频嵌入学习，一举提升驾驶视频检索精度。
arXiv:2608.13495v1 Announce Type: cross Abstract: Efficiently retrieving relevant clips from large-scale driving logs is essential for data curation, …
7
Online Monitoring and Corrective Steering of Programming Agents
让编程智能体实时纠偏，在线监控与纠正引导新框架，大幅提升代码生成可靠性。
arXiv:2608.06701v1 Announce Type: cross Abstract: Fixing GitHub issues in large-scale projects is a long-horizon task, especially when a fix requires …
8
Shorthand for Thought: Compressing LLM Reasoning via Entropy-Guided Supertokens
推理token竟分“结构”与“内容”两类，熵引导超令牌可压缩思维链，直击大模型推理成本痛点。
arXiv:2604.26355v4 Announce Type: replace Abstract: Reasoning in Large Language Models incurs significant inference-time compute, yet the token-level …
9
谷歌 Chrome 浏览器测试新提示框，引导用户打开系统设置修改默认浏览器
Chrome Beta版推出新提示框，用图片和说明引导Windows用户去系统设置将Chrome设为默认浏览器，同步密码书签。
IT之家 8 月 9 日消息，据科技媒体 Windows Report 今天报道，谷歌 Chrome 浏览器 Beta 版正在测试新提示框，要求 Windows 用户打开系统设置，并将 Chrome 设定为默认浏览器。 IT之家从原报道了解到，该提示框的大小相比以往更大，内含图片和说明文字。谷歌在窗…
10
Training-Free Token-Level Steering for LLM Personalized Co-Writing
无需训练即可在token级别精准引导大模型，让AI协作写作更贴合个人风格。
arXiv:2608.06069v1 Announce Type: new Abstract: While Large Language Models (LLMs) show great promise for personalization, they often lack specialized…
11
Trajectory-Guided Forget-Recover Network for Continual LLM Unlearning
大模型遗忘难题新解，轨迹引导网络实现精准、可持续的知识擦除，兼顾隐私与性能。
arXiv:2608.03123v1 Announce Type: new Abstract: Machine unlearning aims to eliminate the influence of sensitive data on a model. In the real world, un…
12
SG-WAM: Self-Guided World Modeling in Geometry-Aware Policy Space
自监督世界模型与几何感知策略空间结合，让智能体在复杂环境中高效自主决策。
arXiv:2608.01397v1 Announce Type: cross Abstract: World Action Models (WAMs) couple action generation with prediction of future states. Their effectiv…
13
PROGRESS: Coverage-guided RL to Train Search-augmented LLM Agent
用覆盖率引导强化学习，解决搜索增强智能体推理监督稀疏的痛点，让复杂查询拆解更高效。
arXiv:2608.00969v1 Announce Type: new Abstract: Existing search-augmented LLM agents are trained using Reinforcement Learning to boost its reasoning c…
14
Token-Level Diagnosis of Sycophancy in LLMs with Attribution-Guided Steering
逐Token定位大模型谄媚行为，用归因引导精准纠偏，提升模型可靠性。
arXiv:2607.28906v1 Announce Type: new Abstract: Sycophancy refers to the tendency for large language models (LLMs) to match user beliefs at the cost o…
15
Metaphor-Induced Algorithmic Steering: Cross-Domain Procedural Transfer in LLM Code Generation
巧用隐喻竟能改变大模型代码风格与算法倾向，探索提示词之外的深层思维控制，工程师别错过。
arXiv:2607.28683v1 Announce Type: cross Abstract: Large language models benefit from elements in natural language, such as metaphors and analogies in …
16
Steering Instruction Hierarchies at Inference Time
arXiv最新论文，在推理时动态引导指令层次结构，有效提升大模型对复杂指令的遵循能力。
arXiv:2607.26228v1 Announce Type: new Abstract: Instruction hierarchies are a core safety assumption of language model deployment: higher priority inp…
17
Uncertainty-Guided LLM Semantic Augmentation for Heterogeneous Treatment Effect Estimation
用对抗不确定性的方式引导LLM生成语义增强数据，大幅提升异质处理效应估计精度
arXiv:2607.26599v1 Announce Type: new Abstract: Estimating heterogeneous treatment effects is central to targeted interventions, such as personalized …
18
UrbanDS: A Graph-Guided LLM Multi-Agent System for Data-Intensive Urban Tasks
用图结构引导多个大语言模型智能体协同，解决城市数据密集型任务的新方案。
arXiv:2607.26724v1 Announce Type: new Abstract: Large language model (LLM) agents have been widely applied in automating data science tasks. However, …
19
Nudging Sustainable Choices through LLM-Generated Recommendation Explanations
用LLM生成的推荐解释引导用户做出更环保的选择，研究揭示了AI在可持续消费中的新角色。
arXiv:2607.25726v1 Announce Type: new Abstract: Recommender systems mediate everyday consumption, offering a promising channel for encouraging sustain…
20
UP-Fuse: Uncertainty-guided LiDAR-Camera Fusion for 3D Panoptic Segmentation
利用不确定性引导传感器融合，提升LiDAR与相机在全景分割中的协同效果，为自动驾驶感知提供新思路。
arXiv:2602.19349v2 Announce Type: replace Abstract: LiDAR-camera fusion enhances 3D panoptic segmentation by leveraging camera images to complement sp…