基准测试的审计本身也可能出错？这篇论文揭示了基准有效性审计的五种典型失败模式，对于构建更可靠的AI评估体系是关键警示。
arXiv:2607.02586v1 Announce Type: new Abstract: Governance frameworks ask AI providers and auditors for documented evaluation evidence, and perturbati…
两套AI商业推荐系统在推荐内容上分歧明显，但诊断出的故障模式却惊人一致，揭示AI推荐逻辑的深层规律。
arXiv:2606.26116v1 Announce Type: cross Abstract: A brand whose customers use both ChatGPT and Claude for product recommendations faces a strategic ch…
揭秘LLM代理在长程任务中过早承诺的失败模式，提供系统诊断方法，提升AI决策可靠性。
arXiv:2606.22936v1 Announce Type: new Abstract: Long-horizon LLM agents can fail quietly: they settle on one reading of the evidence early, then spend…
研究大语言模型在硬件RTL编码中的失败与泛化表现，揭示AI辅助硬件设计的边界与潜力。
arXiv:2606.19347v1 Announce Type: cross Abstract: Translating sequential programming priors into the parallel temporal logic of hardware design remain…
用“周期表”系统梳理LLM推理的范式、方法与失败模式，结构清晰，综述全面，值得研究者细读。
arXiv:2606.11470v1 Announce Type: new Abstract: Large Language Models (LLMs) have achieved strong performance across natural language processing tasks…
新方法用“诱饵校准”精准定位大模型失败模式，避免虚假发现
arXiv:2606.09046v1 Announce Type: new Abstract: Useful audits reveal not only how often a model fails, but also where its failures concentrate. An aud…
系统性揭示LLM在生物与经济安全基准上的类失控优化器失败模式，视角新颖，观察格式简化。
arXiv:2509.02655v3 Announce Type: replace-cross Abstract: Many AI alignment discussions of "runaway optimisation" focus on RL agents: unbounded utilit…
揭示AI研究报告的5种系统性失败模式，帮你揪出单位换算等隐藏错误，提升报告可信度
AI research reports look authoritative. The numbers line up, the charts are clean, and every claim has a source citation. But when you actually open t…
深入剖析vibe coding的适用边界，帮开发者识别何时该回归传统工程方法
Vibe coding is real, useful, and produces working software. It is also, past a certain point, a way of generating a mess faster than any one person ca…
揭秘构建AI编码代理的八大失败模式与支柱，用现有工具组合打造可靠“缰绳”
Teams building with AI usually end up building two products: the thing they ship, and the system around their agents that makes them useful in buildin…
诊断并修复自演化大模型技能库中一种静默失败模式，为提升LLM长期稳定性提供新视角。
arXiv:2605.19576v1 Announce Type: cross Abstract: Self-evolving skill libraries face a silent failure mode we term \emph{library drift}: unbounded ski…