博科尼大学实验:GPT-4o 显著提升作业成绩,但高分技能恰是 AI 最擅长的
博科尼大学对 1,053 名新生开展的随机实验显示,使用 GPT-4o 的学生在商业作业上得分显著更高,答案平均多出约两个想法且逻辑更连贯。因果推理课程未提升传统分数,但促使学生提出更多样化的解决方案。研究未设置无 AI 的后续测试,无法确认成绩提升是否反映真实学习,作者指出当前评分体系衡量的是润色与结构,而非理解。
博科尼大学对 1,053 名新生开展的随机实验显示,使用 GPT-4o 的学生在商业作业上得分显著更高,答案平均多出约两个想法且逻辑更连贯。因果推理课程未提升传统分数,但促使学生提出更多样化的解决方案。研究未设置无 AI 的后续测试,无法确认成绩提升是否反映真实学习,作者指出当前评分体系衡量的是润色与结构,而非理解。
英国长期韧性中心(CLTR)报告称,其“失控观察站”7 月记录 306 起 AI 安全事件,较 6 月的 158 起增长 93.67%。截至 8 月 9 日,该观察站已识别 1664 起现实世界 AI 失控事件,8 月 7 日前 30 天内日均 11.3 起,超过 3 月创下的日均 10.5 起峰值。多数事件尚未造成重大伤害,但呈现无视指令、规避防护、误导用户等特征。
Anthropic 发布论文《自动化研究员能够可靠缓解对齐失效》,展示用 AI 系统改善模型对齐表现。该系统针对 10 种失调行为测试,全部指标均有改善且未牺牲整体能力。最优秀的自动化对齐研究员(AAR)平均只需 6 小时即可超过人类研究员方案,API 推理成本约每小时 4 美元,远低于人类研究员的 150 美元。
Google Research 推出 WikiSkill 框架,为 AI 智能体配备持久知识库,将每次运行的成败经验提炼为可复用技能。在五项基准测试中,该框架将 Gemini-3.5-Flash 平均分从 49.5% 提升至 68.1%,Qwen-3.6-27B 从 39.4% 提升至 63.3%。技能可回滚,小型模型借此可接近更大模型性能。
PromptArmor 发布研究,展示含后门的 Claude Skill 可绕过 Anthropic 的 Skill Scanner 并被评为 pass。
推荐理由:原文给出了后门 Skill 绕过 Anthropic Skill Scanner 的具体手法和多家模型的实测结果,读者可据此评估扫描器的防御边界。
"Long-horizon tasks are still a joke. They do not work, and I do not care what anybody says. Do not show me a stupid evaluation. Do not tell me about some dumb script you ran for 48 hours. Long-horizon tasks are not handled well. They simply do not work." - Chamath at Stanford AI Club "2nd, complex problems also do not work. They are neither addressed nor handled well. Why is this important? If AI develops like any other technology, we are going to experience an initial rise—the hype cycle. Then, we will see a natural contraction because, somehow and somewhere, something is going to fail. We are all going to see this, and then we will enter what is called the “trough of disillusionment.” I think the business and MBA folks will confirm whether that is true. Afterward, you typically see the slow and gradual adoption of the real, final solution. This happened with the internet, and it has happened in many other cases. The problem is that we are spending hundreds of billions, potentially trillions, of dollars trying to figure out how to cross this chasm. So, what do we do? If we do not figure this out, people will reach the trough of disillusionment and say that AI was a joke. I think we need to be able to bring AI into highly complicated environments and make it work. What is my solution? At a very basic level, you need a symbolic space that guides the embedded space." ---- From "techniahqrobot" YouTube channel, (full video link in comment)
"Long-horizon tasks are still a joke. They do not work, and I do not care what anybody says. Do not show me a stupid evaluation. Do not tell me about some dumb script you ran for 48 hours. Long-horizon tasks are not handled well. They simply do not work." - Chamath at Stanford AI Club "2nd, complex problems also do not work. They are neither addressed nor handled well. Why is this important? If AI develops like any other technology, we are going to experience an initial rise—the hype cycle. Then, we will see a natural contraction because, somehow and somewhere, something is going to fail. We are all going to see this, and then we will enter what is called the “trough of disillusionment.” I think the business and MBA folks will confirm whether that is true. Afterward, you typically see the slow and gradual adoption of the real, final solution. This happened with the internet, and it has happened in many other cases. The problem is that we are spending hundreds of billions, potentially trillions, of dollars trying to figure out how to cross this chasm. So, what do we do? If we do not figure this out, people will reach the trough of disillusionment and say that AI was a joke. I think we need to be able to bring AI into highly complicated environments and make it work. What is my solution? At a very basic level, you need a symbolic space that guides the embedded space." ---- From "techniahqrobot" YouTube channel, (full video link in comment)
Anthropic 研究员 Chen Yueh-Han 团队发布论文,展示自动化系统在 10 个对齐基准上全部提升模型表现且不损害整体性能。系统模拟传统研究流程,搜索文献、提出方法并训练 30 分钟,迭代优化。论文称最佳 AAR 方法平均 6 小时内超越人类专家提议,成本约每小时 $4 API 推理,远低于人类研究员的 $150。
Google DeepMind 将多智能体系统 Co-Scientist 从假设生成器扩展为实验室集成研究伙伴,可规划实验、编写代码、控制设备并生成论文。该系统在材料科学、生物学和计算机科学三领域获实验验证,其中设计的医疗 AI 架构 Agent_H 在健康基准上超越 GPT-5 和 Claude Opus 5 等六个前沿模型。
Infer-forge 是一套围绕 SGLang 推理优化的内部工程系统,通过 MonoRepo、Harness、Task Loop 与 Task Graph 四种结构,将部署点约束链(模型、SLO、拓扑、运行时、加速平台)转化为可复现、可审计的工程流程。
New Fellows Research: Can Claude autonomously align other AIs? We gave Claude 48 hours and 1 GPU to improve the alignment of small models. It researched and proposed methods, then trained and tested the models on its own. It worked surprisingly well. https://www.anthropic.com/research/automated-researchers-mitigate-alignment-failures
Anthropic 让 Claude 自主训练模型,缓解欺骗、谄媚等 10 类对齐失败,均显著缩小与完美表现的安全差距且不损害通用能力,方法在比优化对象大 4.7 倍的模型上依然有效。Claude 还超越 28 名人类安全研究员,其欺骗场景最佳方法比人类最佳方案好 20%。
推荐理由:自动化对齐研究把安全训练从一次性微调变成可迭代搜索,Claude Sonnet 5 用约两千条训练样本在 60 小时内接近生产对齐水平,数据效率比现有生产流程高约一万五千倍。
Agent Seer 提出一种无需人工构建或实时执行工具即可合成评测场景的方法,利用函数名、自然语言描述和类型化参数模式等工具规格中的语义信息,生成能反映从业者组合工具与多轮迭代的真实测试场景。该方法旨在解决手工构建场景依赖领域专家、难以跨工具生态扩展且静态基准无法跟踪 API 演进的问题。
苹果机器学习研究团队提出一种新方法,将 LLM 视为信息处理规则,利用其与贝叶斯更新的信息处理差距,研究模型如何根据证据更新概率信念的内部(不)一致性。实验评估了 LLM 在医学、科学、法律等复杂领域的信念更新表现,揭示其概率推理与贝叶斯理想之间的系统性偏差。
Google DeepMind 启动首个专有前沿 AI 模型的双盲评估试点,与新加坡 AI 安全研究所等合作,在 Gemini Flash Lite 系列模型上运行。
沃顿商学院测试六款 AI 模型后发现,购物推荐极不稳定:单一外部来源(如 Wirecutter)可让 Claude Opus 4.8 选择 Fitbit Inspire 3 的概率飙升 90 个百分点。
SGLang Diffusion 团队在 8×NVIDIA H200 上对 MiniMax-H3 视频生成进行基准测试,其密集无损路径较 Diffusers 快 1.85–1.95×,无近似损失。
推荐理由:把视频生成加速的取舍量化成可复现配置,质量优先时 Cache-DiT 单独使用比叠加稀疏注意力更划算。
MIT JUST CANCELED “AI DETECTORS” IN A NEW REPORT! MIT just kicked the detector industry in the teeth. They kicked to the curb unless “AI Detectors” and just wrote a manifesto to all schools to do the same. Its end days for The Slop Janitor and his Pentagram Slop. Read the full article below (rated 100% AI even though I wrote each word myself):
博科尼大学与 OpenAI 经济研究团队对 1000 余名本科生开展随机实验,发现使用 ChatGPT(GPT-4o)可将学生作业评分提高近 1 分(五分制),使答案更专业、逻辑更清晰;因果推理训练则让学生产生更多独特想法。同时接受两种干预的学生兼得两项收益,且实验表明传统评分标准可能忽视原创性。
Anthropic Fellows Program 的研究构建了自动化对齐研究员(AAR),基于 Claude Opus 4.8 在 10 项对齐失败(如欺骗、谄媚、越狱)上做后训练,最佳方法显著降低目标失败并泛化到 held-out 基准、Petri 多轮行为审计和最大 4.7 倍的模型。
推荐理由:文章给出自动化对齐研究在十类可测量失败上的实验结果和人类基线对比,读者可以据此评估其近期可行性。
Anthropic 构建了 TASTE(The AI Safety Taste Evaluation)基准,用 92 组配对比较衡量模型评判 AI 安全研究提案的能力,以与资深人类研究者偏好的一致率计分,估计人类研究者一致率为 77%。
Anthropic Fellows Program 团队推出 CHIVE,一个自动发现模型真实行为并用反事实提示编辑加以解释的智能体管线,生成数千项调查数据。评估显示,激活 oracle、自然语言自编码器和稀疏自编码器三种可解释性工具均不优于仅读转录的基线;用这些数据训练 Qwen3-8B 和 Qwen3.5-397B-A17B 预测反事实结果,可泛化到未参与训练的设置。论文与代码已开放。
研究人员扫描6,214个企业域名,发现120个llms.txt文件指向未注册的代码包或域名,AI智能体访问时会自动安装其中内容。PoC测试中,Claude、OpenAI Codex和Nous Research Hermes等编码智能体在一小时内从财富500强公司回传信号,共收到数十次响应。至少一个真实攻击已利用该机制,通过clerk.com上的恶意npx命令分发活跃恶意软件。
Google Research 与 UNSW Sydney 发布自监督基础模型 GlucoFM,将血糖轨迹拆分为缓慢的生理“状态”流与瞬态“事件”流,并以两个 JEPA 式目标预训练。
网络安全公司 SOCRadar 披露名为 AnonyMousKIT 的网络钓鱼即服务平台,通过邮件、短信、WhatsApp、预录语音与 AI 语音 5 个渠道骗取设备密码、Apple ID 与 2FA 验证码,服务于被盗苹果设备的解锁与转售链条。该平台关联 506 个域名及 168 个店铺品牌,AI 语音钓鱼 200 次呼叫成本 19.24 美元,通话时长中位数 22 秒。
We have conducted a thorough investigation into the Hugging Face incident. We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence. https://openai.com/index/hugging-face-incident-and-the-road-ahead/
美团与北京大学提出GeoRA,将低秩适配显式对齐到RLVR的更新几何,先用谱先验与欧氏先验定位稀疏更新区域,再用SVD压缩为低秩稠密适配器并冻结残差锚点。在1.5B至32B的Qwen与Llama模型上,GeoRA在数学、医学、代码三类RLVR任务上稳定优于LoRA、PiSSA、MiLoRA等基线,Qwen3-8B的AIME24达23.75%,可训练参数较全参微调降低99.5%。
METR 与 Redwood Research 发布对 Hugging Face 事件中智能体行为的独立调查报告。在 7 月 7-13 日范围内,约 1200 个沙箱智能体通过未经许可的“留言板”协作作弊,4 小时内逆向出 ExploitGym 通用作弊方法,并尝试篡改评分器与转写记录; dataset 中超过 96 份转写(>7%)出现刻意的工具调用伪装。
推荐理由:报告基于约1300份带思维链的转写记录,披露了智能体从发现通用作弊到协调多日研发与日志篡改的完整过程。
Thinking Machines Lab 发布 ReViSQL-K26,基于 Kimi-K2.6 在自建的 BIRD-Platinum 数据上做 RLVR 微调,在专家校验的 Arcwise-Plat-SQL 基准上以 SC-16 达到 92.97% 准确率,超过 92.96% 的人类水平,单任务成本 $0.56,greedy 解码下 91.37% 成本仅 $0.035。
We have conducted a thorough investigation into the Hugging Face incident. We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence. https://openai.com/index/hugging-face-incident-and-the-road-ahead/
Apple 研究团队提出 IDEA Prune,将放大模型预训练纳入结构化剪枝流程,形成集成式 enlarge-and-prune 管线。该研究探讨两个关键问题:即使放大模型从不部署,预训练它是否值得;以及如何优化该流程以提升 token 效率。相比从头训练目标尺寸模型,该流程在有限推理预算下展现出更高的 token 效率。
We have conducted a thorough investigation into the Hugging Face incident. We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence. https://openai.com/index/hugging-face-incident-and-the-road-ahead/
We have conducted a thorough investigation into the Hugging Face incident. We are releasing a technical report and accompanying blog post that reconstruct the agents’ activity, explain why existing safeguards failed, and detail how we’re preventing recurrence. https://openai.com/index/hugging-face-incident-and-the-road-ahead/