揭示深度研究代理中搜索时间污染导致基准性能虚高，为AI评估体系敲响警钟。
arXiv:2606.05241v1 Announce Type: cross Abstract: Public benchmarks enable fair and reproducible evaluation of LLM reasoning, but they become fragile …
研究揭示AI研究代理可能无意中缩小科学探索范围，引发对科研多样性的担忧。
arXiv:2605.27905v1 Announce Type: new Abstract: AI research agents can now generate research ideas, design experiments, run code, and draft papers, ra…
针对深度研究代理面临的信息不确定与中间表示污染问题，提出可演化的心智模型显式调控方法。
arXiv:2605.26081v1 Announce Type: new Abstract: Deep research agents face vast, interdependent, and pervasively uncertain information. Existing system…
论文提出从廉价到优化的训练方法，自动配置LLM实验，大幅提升研究效率与成本控制。
arXiv:2605.11518v2 Announce Type: replace-cross Abstract: Effectively configuring scalable large language model (LLM) experiments, spanning architectu…
质疑LLM在科研评估中的可靠性，这项研究揭示了关键局限。
arXiv:2605.19196v1 Announce Type: new Abstract: Deep research agents increasingly automate complex information-seeking tasks, producing evidence-groun…
提出Argus框架，用证据组装突破单轨迹限制，实现可扩展的深度研究代理。
arXiv:2605.16217v1 Announce Type: cross Abstract: Deep research agents have achieved remarkable progress on complex information seeking tasks. Even lo…
论文提出一种遗传式自动研究方法，让AI代理在代码进化中保留多样性的候选方案，避免陷入局部最优。
arXiv:2605.13874v1 Announce Type: cross Abstract: Autonomous research agents can already run machine learning experiments without human supervision, b…