1
阿里斩获国际AI顶会最佳资源论文奖，提出Agent评测新范式
阿里摘得ACL 2026最佳资源论文奖，以Agent评测新范式突破AI前沿，国内独一份。
2
WildIFEval: Instruction Following in the Wild
新基准WildIFEval带你从实验室走进真实世界的指令跟随能力评估
arXiv:2503.06573v3 Announce Type: replace-cross Abstract: Recent LLMs have shown remarkable success in following user instructions, yet handling instr…
3
What Really Matters for Table LLMs? A Meta-Evaluation of Model and Data Effects
元评估揭示哪些因素真正决定表格大模型性能，模型架构vs训练数据的深度对比。
arXiv:2501.14717v2 Announce Type: replace Abstract: Table modeling has progressed for decades. In this work, we revisit this trajectory and highlight …
4
HarDBench: A Benchmark for Draft-Based Co-Authoring Jailbreak Attacks for Safe Human-LLM Collaborative Writing
人机协作写作竟藏越狱风险？新基准揭示大模型安全新盲区
arXiv:2604.19274v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as co-authors in collaborative writing, where u…
5
Multilingual Refusal Alignment for Safer Large Language Models
从单语到多语：这篇ACL 2026论文教你如何让大模型在多种语言下都学会安全拒答
arXiv:2606.07535v1 Announce Type: new Abstract: As Large Language Models (LLMs) are deployed globally, ensuring their safety and alignment across mult…
6
Self-Consistency from Only Two Samples: CoT-PoT Ensembling for Efficient LLM Reasoning
仅用两个推理样本即可实现LLM自一致性？CoT+PoT集成方案，大幅提升推理效率的新突破。
arXiv:2604.17433v2 Announce Type: replace-cross Abstract: Self-consistency (SC) is a popular technique for improving the reasoning accuracy of large l…
7
Jailbreaking Multimodal Large Language Models using Multi-Clip Video
利用多片段视频时序特性实现多模态大模型攻击，ACL 2026会议研究揭示新安全漏洞。
arXiv:2606.02111v1 Announce Type: cross Abstract: As multimodal large language models (MLLMs) have advanced to process video inputs, concerns have eme…
8
Speculative Decoding Across Languages
arXiv最新论文提出跨语言推测解码方法，无需目标语言训练数据即可加速多语言大模型推理
arXiv:2605.30580v1 Announce Type: cross Abstract: Speculative decoding has become a crucial component of large language model (LLM) inference, enablin…
9
LLMSurgeon: Diagnosing Data Mixture of Large Language Models
诊断大模型数据配比的“DNA”，ACL 2026主会论文开源代码
arXiv:2605.30348v1 Announce Type: cross Abstract: The pretraining data mixture of Large Language Models (LLMs) constitutes their "digital DNA", shapin…
10
TIGER: Text-Informed Generalized Enzyme-Reaction Retrieval
用文本信息检索酶反应，提出通用化框架TIGER，被ACL 2026接收，生物与NLP交叉新范式。
arXiv:2605.24489v1 Announce Type: new Abstract: Enzyme-reaction retrieval is a fundamental problem in computational biology, underpinning enzyme chara…
11
Automated Knowledge Component Generation for Interpretable Knowledge Tracing in Coding Problems
自动生成知识组件，让编程知识追踪更透明，从ACL论文看教育AI新突破。
arXiv:2502.18632v4 Announce Type: replace-cross Abstract: Knowledge components (KCs) mapped to problems help model student learning, tracking their ma…
12
Scaling Laws for Code: A More Data-Hungry Regime
重磅研究：代码领域的缩放定律显示需要比自然语言多几个数量级的数据才能达到相同性能提升，引发对大模型训练数据效率的重新思考。
arXiv:2510.08702v2 Announce Type: replace Abstract: Code Large Language Models (LLMs) are revolutionizing software engineering. However, scaling laws …