通过合成指令数据扩展预训练规模，突破传统监督训练数据瓶颈的新方法
arXiv:2601.22146v2 Announce Type: replace-cross Abstract: Due to limited supervised training data, large language models (LLMs) are typically pre-trai…
从粗到细监督实现高效高分辨率视觉表示，SigLIP-HD在处理多模态大模型时兼顾计算效率与细粒度特征。
arXiv:2607.09488v1 Announce Type: new Abstract: High-quality visual representation is a long-standing pursuit in computer vision. In the context of mu…
自我引导的测试时训练，专治长上下文LLM的准确率退化难题。
arXiv:2607.09415v1 Announce Type: cross Abstract: Long-context processing has become increasingly important for large language models (LLMs), but simp…
DeepMind的AlphaZero在稀疏奖励游戏中表现有限，论文提出辅助监督方法提升性能。
arXiv:2607.08984v1 Announce Type: cross Abstract: AlphaZero has demonstrated that a neural-guided Monte Carlo Tree Search can achieve superhuman perfo…
据韩国金融监督院的声明，其负责人呼吁该国资产管理公司对ETF产品承担更大的责任。院长李粲珍周一会见了20家资产管理公司的首席执行官和韩国金融投资协会的负责人，李粲珍表示，随着ETF市场快速扩张，资产管理公司的责任也日益加重，他呼吁这些公司在宣传此类产品时确保投资信息的准确性。李粲珍还提醒资产管理公司…
利用激活几何进行无监督特征挖掘，无需人类标注即可揭示大模型内部表征
arXiv:2607.04222v1 Announce Type: new Abstract: Interpretability methods aim to reveal the features represented inside large language models (LLMs). M…
无需人工标注，通过神经元激活模式筛选数据，实现LLM高效自蒸馏训练。
arXiv:2607.02460v1 Announce Type: cross Abstract: Post-training large language models (LLMs) without real-world interaction feedback or human-labeled …
提出QVal框架，低成本评估长程LLM代理的密集监督信号，提升任务效率。
arXiv:2606.32034v1 Announce Type: cross Abstract: LLM agents increasingly act over long horizons, where a single trajectory can contain hundreds or th…
不用标注数据也能精准识别漏洞？ANVIL利用LLM的异常检测新范式，突破监督学习局限
arXiv:2408.16028v4 Announce Type: replace-cross Abstract: Supervised-learning-based vulnerability detectors often fall short due to limited labelled t…
一篇挑战LLM后训练中在线模仿学习优势的论文，深入剖析了非可实现性与时间跨度的关键作用
arXiv:2606.30445v1 Announce Type: new Abstract: Online imitation learning (IL), particularly on-policy distillation, has emerged as a strong LLM post-…
AI编程助手Cursor让开发者放手，无人工审查也能直接上线生产环境，代码生成信任度飙升
IT之家 6 月 29 日消息，据 Business Insider 报道，AI 编码智能体正越来越多地获得信任，可在无人工监督的情况下独立完成工作。 人工智能编程助手 Cursor 的最新数据显示，过去半年里，无需额外人工审核、直接上线生产环境的 AI 生成代码变更占比大幅飙升。 这一趋势表明，开…
近日，市场监管总局起草形成《实施价格监督员制度暂行规定（征求意见稿）》，自6月29日至7月28日向社会公开征求意见。市场监管总局表示，建立价格监督员制度，是深入整治“内卷式”竞争的有力抓手，是完善价格治理机制的制度创新，是规范市场价格秩序的重要途径。有利于压实企业价格合规管理主体责任，健全行业协会自…
首次验证无标准答案的强化学习也能提升LLM，颠覆传统监督范式。
arXiv:2606.27369v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) for training LLMs typically rely on ground-truth…
混合LLM代理实现通用指南驱动的图像聚类，突破传统方法对特定数据的依赖。
arXiv:2606.24094v1 Announce Type: new Abstract: Unifying image clustering across different clustering scenarios remains challenging due to fundamental…
结合Hessian矩阵增强监督学习，高效求解高维HJB偏微分方程，AI+数值方法新进展。
arXiv:2606.23827v1 Announce Type: cross Abstract: A data-driven method is developed for approximating value functions in deterministic optimal control…
挑战测试时训练新范式：用语言模型自监督的下一词预测信号作为内循环目标，无需额外损失函数。
arXiv:2606.21803v1 Announce Type: new Abstract: Next-token prediction is the self-supervised signal that trains language models, and every observed pr…
应对LLM后训练中噪声监督，VRPO提出重新思考价值建模实现鲁棒强化学习，值得关注。
arXiv:2508.03058v2 Announce Type: replace-cross Abstract: Reinforcement Learning (RL) in real-world environments often suffers from ambiguous or incom…
LLM代理结合PDDL规划，端到端有监督训练编排器，突破传统规划瓶颈。
arXiv:2606.21740v1 Announce Type: new Abstract: Translating natural-language planning intent into verified plans is a longstanding challenge: people c…
从海量网络数据中精准提取有效信号，创新医疗编码器预训练的数据配方，为专业领域模型训练提供新思路。
arXiv:2606.22079v1 Announce Type: cross Abstract: Web data curation has been widely studied for decoder Large Language Model (LLM) pretraining. Encode…