1
On the Role of Citations in Preference Data
一篇探讨引用在偏好数据中作用的学术论文，为AI对齐与偏好学习研究提供新视角。
arXiv:2608.21376v1 Announce Type: cross Abstract: Many NLP tasks require systems to provide attribution in their outputs--i.e. citations to grounding …
2
Google Discover is getting an AI chatbot-tuned feed
AI 根据聊天偏好实时调整内容流，记住你的喜好，让每次刷新都更懂你。
Google will soon allow you to customize your Discover feed by describing what you want to see. The new feature, rolling out to the Google app in the "…
3
Self- and Other-Labels Induce Bidirectional Bias in LLM Judges
LLM裁判也有“偏心眼”：自我标签与他者标签会引发双向评估偏差，AI评测可信度再遭挑战。
arXiv:2608.18091v1 Announce Type: cross Abstract: As LLM-as-a-judge systems become increasingly widespread, self-preference in LLMs -- the tendency to…
4
Data-DPO: Direct Preference Optimization for Target Model Data Selection in LLM Post-Training
数据选择新思路：用DPO偏好优化为模型量身筛选训练数据，提升后训练效果。
arXiv:2608.16926v1 Announce Type: new Abstract: Data selection in supervised fine-tuning aims to select a small set of effective samples from large-sc…
5
Leveraging Machine Unlearning for Cost-Efficient Preference Alignment
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…
6
Learning to Adapt Cross-Domain Preferences via Meta-LoRA for LLM Personalization
元学习+LoRA让大模型快速适应跨域偏好，个性化调校从此更聪明高效。
arXiv:2608.12389v1 Announce Type: new Abstract: Cross-domain zero- or few-shot personalization aims to generate user-preferred responses in unseen con…
7
Overconfident and Blind to Details: Fixing Prompt Insensitivity with Abductive Preference Learning
大模型竟对五条腿的狗视而不见？这项研究用溯因偏好学习破解提示不敏感难题。
arXiv:2510.09887v3 Announce Type: replace Abstract: Vision and language models frequently ignore semantically critical input edits, defaulting to pret…
8
Procedural Fairness Failures in RLHF from Preference Averaging
从偏好平均视角揭示RLHF中程序公平性缺陷，为对齐公平研究提供新警示。
arXiv:2608.10126v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single r…
9
Uncertainty-Aware Variational Reward Factorization via Probabilistic Preference Bases for LLM Personalization
把个性化奖励拆解成可解释概率偏好，不确定性建模让LLM对齐更稳更准。
arXiv:2604.00997v2 Announce Type: replace Abstract: Reward factorization personalizes large language models (LLMs) by decomposing rewards into shared …
10
Cross-Model Humor Preference Modeling with Cards Against Humanity
用Cards Against Humanity测试大模型间的幽默偏好，看GPT-4o能否猜中Claude的笑点。
arXiv:2608.07481v1 Announce Type: cross Abstract: This paper investigates whether one large language model can approximate the humor preferences of an…
11
LangChoiceBench: Measuring and Explaining Programming-Language Choice in LLMs
用全新基准测试大模型选编程语言的偏好，看清AI写代码时的“语言心思”
arXiv:2608.06041v1 Announce Type: cross Abstract: Large language models (LLMs) have been shown to exhibit strong Python preferences when generating pr…
12
Preferred, Not Safer: Pairwise Preference Is a Poor Proxy for Clinical Safety
临床医生的成对偏好并不等于安全，这项研究给AI医疗评估敲响警钟
arXiv:2608.02617v1 Announce Type: cross Abstract: We evaluate whether clinician pairwise preferences provide a reliable signal of clinical safety in l…
13
Style Wins, Substance Loses: A Diagnosis of LLM-as-Judge in Idea Generation
揭示LLM评审在创意评估中重风格轻实质的偏见，挑战AI评价可靠性。
arXiv:2608.01666v1 Announce Type: cross Abstract: However, whether these judges truly evaluate the scientific substance of ideas or are influenced by …
14
Show HN: Offline Flock Navigation
离线导航新选择，路线自动绕开摄像头，全程无云端参与，隐私至上的驾驶助手。
Hello HN! A little side project I have been working on a fork of OsmAnd with offline ALPR/Flock camera avoidance routing. Also, took the time to at le…
15
RM-Distiller: Exploiting Generative LLM for Reward Model Distillation
生成式大模型蒸馏奖励模型新方法，解决偏好标注昂贵难题。
arXiv:2601.14032v2 Announce Type: replace Abstract: Reward models (RMs) play a pivotal role in aligning large language models (LLMs) with human prefer…
16
Co-design of LLM-based preference agents: participation may drive overtrust
参与式设计AI偏好智能体可能引发过度信任，需警惕用户对LLM代理的盲目依赖。
arXiv:2607.21757v1 Announce Type: cross Abstract: Large language models are increasingly used to simulate human preferences in research and practical …
17
Fragile Preferences: A Deep Dive Into Order Effects in Large Language Models
揭秘大模型如何被输入顺序“带偏”——这项研究深入分析了LLM偏好的脆弱性，对理解模型鲁棒性至关重要。
arXiv:2506.14092v4 Announce Type: replace Abstract: Large language models (LLMs) are increasingly deployed in decision-support systems for high-stakes…
18
Rewarding Better Thinking for LLM Preference Alignment
提出通过奖励更优思考过程来对齐LLM偏好，超越仅评估最终结果，提升推理轨迹指导。
arXiv:2607.19824v1 Announce Type: cross Abstract: LLM preference alignment aims to optimize models toward human preferences across diverse user instru…
19
RIMS: Preference Optimization via Smoothed Multi-pair Aggregation for Small-Scale LLM Retrieval-Augmented Generation
提出RIMS方法，通过平滑多对聚合优化偏好，让小规模LLM在RAG任务中性能飙升。
arXiv:2607.16431v1 Announce Type: new Abstract: Small-scale language models (SLMs) are attractive for retrieval-augmented generation (RAG) in resource…
20
Normalized Rewards for Preference Optimization
提出归一化奖励方法，提升偏好优化训练稳定性与效果
arXiv:2607.16240v1 Announce Type: cross Abstract: Direct Alignment Algorithms (DAAs) such as DPO have become a common way to post-train and align LLMs…