LLM代理能否智能地“不作为”？新研究探索Agent安全决策边界，为AI可靠性提供关键视角。
arXiv:2607.10059v1 Announce Type: new Abstract: Agent systems based on large language models (LLMs) are increasingly deployed for autonomous tasks, ye…
顶级AI学者论文直指：基于大语言模型的人类模拟仍不可靠，结论值得关注。
arXiv:2501.08579v3 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly employed for simulating human behaviors across diver…
AI成功融入劳动力需要经济价值作为桥梁，这篇论文探讨人机协调与合作的底层逻辑。
arXiv:2607.03181v1 Announce Type: cross Abstract: Successful diffusion of AI in the workforce hinges on the economic value that AI brings to human end…
探索LLM如何模拟人类群体智能，突破规模与协调限制，带来更优集体决策
arXiv:2606.31404v1 Announce Type: new Abstract: Human swarm intelligence demonstrates remarkable collective accuracy but faces scalability constraints…
新论文提出人类创造力基准，为评估AI与人类创造力提供量化标尺，推动创造性AI研究。
arXiv:2606.30561v1 Announce Type: cross Abstract: Modern AI evaluation frameworks treat evaluator disagreement as noise to be resolved. In creative do…
OpenAI最新研究，基于Codex实验揭示AI从工具向自主智能体转变的关键证据。
Article URL: https://cdn.openai.com/pdf/5d1e1489-21c0-43e4-9d42-f87efdbf0082/the-shift-to-agentic-ai-evidence-from-codex.pdf Comments URL: https://new…
OpenAI最新研究揭示，AI代理正通过处理更复杂、更持久的任务，全面重塑工作模式与生产力边界。
A new OpenAI research paper shows how AI agents are transforming work, enabling longer, more complex tasks and expanding productivity across roles.
新方法！假设驱动优化LLM智能体技能，显著提升复杂任务适应能力
arXiv:2606.22330v1 Announce Type: new Abstract: External skills can improve action-oriented LLM agents without changing model weights, but persistent …
用Claude把研究论文变成临时领域专家，这个工作流太聪明了。
One workflow I like: collecting high-quality research papers on a topic, uploading them to Claude, and turning it into a temporary expert on that doma…
用大模型自动生成代码文档，并引入多评判机制评估质量，提升开发效率
arXiv:2606.09852v1 Announce Type: cross Abstract: High-quality source code documentation is vital yet often neglected, especially in critical domains …
2026年上半年必看LLM研究论文清单，由机器学习专家精心筛选，帮你省时高效追踪前沿进展。
Article URL: https://magazine.sebastianraschka.com/p/llm-research-papers-2026-part1 Comments URL: https://news.ycombinator.com/item?id=48446264 Points…
大模型在复杂语境下会「失忆」，系统揭示 LLM 上下文不变性的严重缺陷与根源
arXiv:2603.23485v2 Announce Type: replace-cross Abstract: Standard evaluation practices assume that large language model (LLM) outputs are stable when…
用LLM智能体打通时间序列预测的“最后一公里”，前沿方法探索。
arXiv:2606.02497v1 Announce Type: new Abstract: Time series forecasting has advanced rapidly, especially with the emergence of foundation models that …
多智能体系统新框架AgensFlow，提出协调策略基底，为复杂AI协作提供理论支持。
arXiv:2605.27466v1 Announce Type: cross Abstract: Multi-agent systems built on large language models (LLMs) require many coordination choices that are…
质疑LLM是否具备真正的内省能力，一项基于实证的严谨检验。
arXiv:2605.26242v1 Announce Type: new Abstract: Can large language models detect and report their own internal states? A number of studies have argued…
LLM代理自我进化新突破：协同策略与环境共同进化，解决代理-环境错配难题
arXiv:2605.24426v1 Announce Type: new Abstract: Large Language Model (LLM) agents are increasingly improved through interaction, yet most self-evoluti…
利用Claude和研究论文，解决实时语音翻译中的性别识别难题，AI辅助编程的实战案例。
Introduction When building apps with AI-assisted coding, you get to decide "what to build." You can set the design and the policy yourself. However, y…
用AI重新定义科研论文的阅读与检索方式，一篇探讨交互界面创新的思考。
Article URL: https://justinross.substack.com/p/an-ai-interface-for-research-papers Comments URL: https://news.ycombinator.com/item?id=48269299 Points:…
从研究论文看GitHub上的开发者行为，揭示面试中难以言说的真实一面
When I started working on my usability study project with KServe, I interacted with KServe users to understand the challenges they were experiencing w…
提出逐步评分奖励机制，优化LLM推理的中间步骤监督，突破传统仅奖励最终答案的局限。
arXiv:2605.17291v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) is widely used to improve reasoning in large lan…