LLM作为评判者自身结果不一致？揭秘模型版本漂移与提示模糊性的根源与解法
Same outputs, same judge, two runs, two scores. The gate flickered red then green on a branch with zero code changes, and that flapping cost me more t…
探寻语言模型内部语义锚定机制，精准定位关键路径，推动模型可解释性研究。
arXiv:2606.12818v1 Announce Type: cross Abstract: Irrelevant numbers in a prompt can shift language model judgments, producing anchoring effects in nu…
注意力机制为LLM推理提供“预规划-锚定”节奏，突破强化学习的粗放信用分配，实现细粒度策略优化。
arXiv:2510.13554v2 Announce Type: replace-cross Abstract: The reasoning pattern of Large language models (LLMs) remains opaque, and reinforcement lear…
一种新记忆进化机制eMoT，通过符号锚定与记忆腐蚀，在数学推理任务中实现17.6%准确率提升，Game of 24达100%。
arXiv:2606.02054v1 Announce Type: new Abstract: While Large Language Models (LLMs) achieve impressive performance on multi-step reasoning tasks, their…
新方法破解LLM事实幻觉：推理时解耦注意力路由与信号增益，实现上下文共振锚定
arXiv:2606.01923v1 Announce Type: cross Abstract: Large Language Models (LLMs) frequently exhibit "contextual disregard" when faced with input evidenc…
提出基于锚定token级logits的LLM越狱检测方法SelfGrader，高效识别对抗性攻击。
arXiv:2604.01473v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are powerful tools for answering user queries, yet they remain …
挑战主流观点：主动代理或许不需要LLM来决策唤醒与锚定，这篇论文带来全新视角。
arXiv:2605.30152v1 Announce Type: cross Abstract: Proactive agents read user activity as text and call an LLM on every event to decide whether to act.…
提出MIRA方法，通过中训练评分锚定实现来源感知数据选择，提升大模型训练质量。
arXiv:2605.30288v1 Announce Type: new Abstract: Mid-training has become an important stage in modern LLM development, using large-scale curated mixtur…
将大语言模型与原子锚定技术结合，演示AI在化学逆合成中的精准应用。
arXiv:2510.16590v2 Announce Type: replace-cross Abstract: Applications of machine learning in chemistry are often limited by the scarcity and expense …
提出专家引导的后合并量化方法，利用合并权重锚定，在低资源部署中平衡模型压缩与性能。
arXiv:2605.16882v1 Announce Type: new Abstract: Low-resource deployment constraints have made model quantization essential for deploying neural networ…
用方程锚定工具使用，让多模态大模型实现相机鲁棒的3D定位，突破视觉视角限制。
arXiv:2605.19528v1 Announce Type: new Abstract: 3D localization in Multimodal Large Language Models (MLLMs), including 3D object detection and 3D visu…