1
Standard Language Ideology in AI-Generated Language
揭露AI生成语言暗中强化标准语言意识形态，关乎语言多样性与公平性
arXiv:2406.08726v3 Announce Type: replace Abstract: Large language models (LLMs) generate text that reinforces standard language ideology: a bias towa…
2
Conditional Vendi Score: Prompt-Aware Diversity Evaluation for Generative AI Models and LLMs
提出条件Vendi评分，首个能结合提示上下文评估生成模型多样性的方法，精准衡量AI创造力。
arXiv:2411.02817v2 Announce Type: replace-cross Abstract: Generative models guided by text prompts are widely evaluated for fidelity and prompt alignm…
3
Scaling Participation in Modular AI Systems
直指当前LLM集中化弊端，提出模块化参与机制以拥抱人类智能多样性。
arXiv:2606.07812v1 Announce Type: cross Abstract: Humanity is a mosaic of multifaceted talents and needs, and any truly intelligent AI must reflect th…
4
Recovering Diversity Without Losing Alignment: A DPO Recipe for Post-Trained LLMs
为LLM后训练设计新方法REDIPO，在不损失对齐的前提下恢复输出多样性，解决指令回答单一化问题。
arXiv:2605.30021v2 Announce Type: replace Abstract: Many open-ended instructions have multiple valid answers that users can benefit from seeing, but p…
5
A Registry-Bound LLM Pipeline for Evidence-Grounded Trait Extraction across Tropical Plants, Aquatic Species, and Exotic Pets
一种注册绑定的LLM管道，为热带植物、水生动物及异宠的性状提取提供可溯源证据，方法论严谨且跨物种适用。
arXiv:2606.00994v1 Announce Type: new Abstract: We describe a registry-bound large-language-model extraction pipeline producing evidence-grounded stru…
6
Meta平台全球升级青少年内容管控，Instagram上线新工具优化推送内容多样性
Meta为青少年升级内容管控，Instagram新工具让推送更丰富多元。
Meta周二宣布，将在全球范围内升级Instagram、脸书（Facebook）及聊天软件Messenger的青少年账号内容管控设置，保障未成年用户浏览适配其年龄段的内容体验。当下各大社交媒体平台因未成年人身心健康与使用安全问题，正受到监管层面愈发严苛的审查。（新浪财经）
7
Anchorless Diversification for Parallel LLM Ideation
提出无锚点多样化策略，提升大模型并行创意生成的质量与多样性，为AI创意优化提供新思路。
arXiv:2605.30150v1 Announce Type: new Abstract: LLMs are increasingly used to generate candidate-idea pools for creative tasks where broad exploration…
8
Quality-constrained Entropy Maximization Policy Optimization for LLM Diversity
提出保证生成质量的同时最大化输出多样性的策略优化方法，破解LLM内容同质化难题。
arXiv:2602.15894v2 Announce Type: replace-cross Abstract: In many large language model (LLM) alignment applications, users expect not only high-qualit…
9
Elias in the Lighthouse, Again? Diagnosing Low Diversity in LLM Stories
通过2万条故事样本，发现大模型输出中11个词出现率高达88.3%，揭示故事生成的同质化顽疾
arXiv:2605.26492v1 Announce Type: cross Abstract: LLM-generated stories are a popular use case, but they show very low variability. We sample 20,000 t…
10
Balancing Uncertainty and Diversity of Samples: Leveraging Diversity of Least, High Confidence Samples for Effective Active Learning
提出Leveraging Diversity of Least方法，通过平衡不确定性与多样性优化主动学习样本选择，提升模型训练效率
arXiv:2605.22169v1 Announce Type: new Abstract: Deep learning models, including Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), h…
11
Vector Policy Optimization: Training for Diversity Improves Test-Time Search
策略优化新范式：训练阶段注入多样性，显著提升测试时搜索性能
arXiv:2605.22817v1 Announce Type: cross Abstract: Language models must now generalize out of the box to novel environments and work inside inference-s…
12
Weasel: Out-of-Domain Generalization for Web Agents via Importance-Diversity Data Selection
ICML 2026新方法：用重要性-多样性数据选择提升Web智能体跨域泛化能力
arXiv:2605.20291v1 Announce Type: new Abstract: Large language models (LLMs) have enabled web agents that follow natural language goals through multi-…
13
Mask-to-Correct$^+$: Leveraging Retriever Diversity for Masking-guided Faithful Fact Correction
利用检索器多样性提升掩码引导事实校正的忠实性，高效融合多源证据。
arXiv:2605.18776v1 Announce Type: cross Abstract: The rapid spread of misinformation on social media highlights the need for robust, automated fact co…
14
Airborne DNA can yield insights with the right techniques
空气DNA采样技术革新生物多样性监测，Nature揭示关键方法细节与前沿应用
Nature, Published online: 19 May 2026; doi:10.1038/d41586-026-01604-7 Airborne DNA can yield insights with the right techniques
15
The Homogenization Problem in LLMs: Towards Meaningful Diversity in AI Safety
大模型同质化如何威胁AI安全？这篇论文系统剖析了症结与出路。
arXiv:2601.06116v5 Announce Type: replace-cross Abstract: Generative AI models reproduce the human biases in their training data and further amplify t…
16
PQR: A Framework to Generate Diverse and Realistic User Queries that Elicit QA Agent Failures
提出PQR框架自动生成多样真实用户查询，精准发现QA agent的失败边界，补足对抗性测试的盲区
arXiv:2605.16551v1 Announce Type: new Abstract: Evaluating LLM-based agents remains challenging because identifying meaningful failure cases often req…
17
Leveraging Error Diversity in Group Rollouts for Reinforcement Learning
强化学习新方法：利用群组展开中的误差多样性提升性能，理论分析与实验验证兼具。
arXiv:2605.17333v1 Announce Type: new Abstract: Reinforcement Learning from Verifiable Rewards (RLVR) typically samples multiple responses per prompt …
18
SEED: Targeted Data Selection by Weighted Independent Set
由相似图构建加权独立集，平衡样本质量与多样性，为高效数据选择提供新框架。
arXiv:2605.15691v1 Announce Type: new Abstract: Data selection seeks to identify a compact yet informative subset from large-scale training corpora, b…
19
Agreement, Diversity, and Polarization Indices for Approval Elections
批准选举中衡量共识、多样性与极化的创新指数，数学严谨且应用潜力大
arXiv:2605.14983v1 Announce Type: cross Abstract: An index is a function that given an election outputs a value between 0 and 1, indicating the extent…
20
GEAR: Genetic AutoResearch for Agentic Code Evolution
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…