用机器遗忘替代昂贵的人类反馈，低成本实现大模型偏好对齐，ICML 2026新思路。
arXiv:2504.06659v2 Announce Type: replace-cross Abstract: Despite advances in Preference Alignment (PA) for Large Language Models (LLMs), mainstream m…
深度学习中模型学到的是真实生物信号还是伪影？ICML 2026 论文以生物成像为例，拷问表征学习的本质，值得每一位科研工作者反思。
arXiv:2603.13377v2 Announce Type: replace-cross Abstract: Representation learning has driven major advances in natural image analysis by enabling mode…
表格数据基础模型如何兼顾公平性？ICML spotlight论文提出全新训练思路，AI公平性研究者必读。
arXiv:2608.14211v1 Announce Type: cross Abstract: Tabular Foundation Models (TFMs) have emerged as leading methods for tabular predictive tasks, lever…
用谱视角剖析大模型领域适应，Diffract方法为持续预训练提供新洞察。
arXiv:2608.10850v1 Announce Type: new Abstract: We study continual pre-training (CPT) as a mechanism for adapting general-purpose large language model…
用模型差分法定位多模态模型内部特征，实现精准控制与可解释性，干货满满。
arXiv:2608.09928v1 Announce Type: cross Abstract: Multimodal Large Language Models (MLLMs) exhibit strong visual understanding, yet the internal featu…
将LLM工作流建模为概率程序，ICML 2026新方法让复杂AI流程更严谨可控。
arXiv:2608.05234v1 Announce Type: new Abstract: Building reliable applications that leverage large language models (LLMs) remains a significant challe…
稀疏注意力中因果信息流被重新定义，Stem提出新视角，为长上下文模型效率与准确性兼得提供关键思路。
arXiv:2603.06274v2 Announce Type: replace-cross Abstract: The quadratic computational complexity of self-attention remains a fundamental bottleneck fo…
提出多任务GRPO方法，有效提升大模型跨任务推理的可靠性，已被ICML 2026收录。
arXiv:2602.05547v2 Announce Type: replace-cross Abstract: RL-based post-training with GRPO is widely used to improve large language models on individu…
注入随机噪声增强LLM推理多样性，ICML 2026 Spotlight论文新思路
arXiv:2605.11936v2 Announce Type: replace Abstract: Recent soft prompt research has tried to improve reasoning by inserting trained vectors into LLM i…
用生物动力学知识先验解决数据稀缺难题，为生物过程建模开辟新路径——ICML 2026 AI for Science 论文
arXiv:2607.20539v1 Announce Type: cross Abstract: While deep learning has accelerated drug discovery, its impact on biomanufacturing has been consider…
ICML'26提出概念集中方法，通过忠实表示干预提升AI可解释性，为模型控制提供新思路。
arXiv:2505.18672v2 Announce Type: replace Abstract: Representation intervention aims to localize and modify the representations that encode the underl…
新方法通过奖励感知群体缩放优化进化策略，显著提升LLM微调效率，已被ICML 2026 workshop收录。
arXiv:2607.19408v1 Announce Type: new Abstract: Using Evolutionary Strategies (ES) for fine-tuning large language models is attractive because it is m…
提出可证明安全的高效聚类方案，为大规模LLM推理保驾护航。
arXiv:2607.19704v1 Announce Type: new Abstract: Scaling LLM-based applications to millions of users is bottlenecked by the inference cost and latency …
提出利用事后经验将结果转化为行动，显著提升长周期语言智能体训练效果，ICML 2026录用论文。
arXiv:2607.16257v1 Announce Type: cross Abstract: Reinforcement learning (RL) has become a widely adopted technique for improving large language model…
低秩预训练大模型遭遇不稳定困境？这项ICML 2026研究提出原生低秩LLM预训练的稳定化方法，兼顾效率与质量。
arXiv:2602.12429v2 Announce Type: replace Abstract: Foundation models have achieved remarkable success, yet their growing parameter counts pose signif…
用行为一致性作为新指标，量化LLM智能体自身不确定性，方法新颖且具启发性。
arXiv:2602.11619v2 Announce Type: replace Abstract: Running the same LLM agent on identical inputs yields 2.3-4.2 distinct action sequences per 10 run…
ICML 2026重磅论文：进化策略替代强化学习，开创大模型微调新范式。
arXiv:2509.24372v3 Announce Type: replace Abstract: Fine-tuning large language models (LLMs) for downstream tasks is an essential stage of modern AI d…
专家训练时长如何影响LLM模型合并效果？这篇ICML 2026 workshop论文揭示关键发现。
arXiv:2607.11997v1 Announce Type: new Abstract: Multi-task model merging combines separately trained expert models into a single model that handles al…
揭秘大模型安全探针为何在最终token失效：早于最后一层的内部表征已暴露漏洞
arXiv:2605.12726v2 Announce Type: replace Abstract: Final-token safety probes monitor a single hidden state after prompt prefill, but jailbreak prompt…
SageMath增强LLM智能体，一场计算与实验数学的自动化探索，看AI如何攻克严谨数学推理。
arXiv:2607.06820v1 Announce Type: new Abstract: Recent advances in AI for Mathematics have focused largely on autoformalization and theorem proving, l…