1
OpenAI ChatGPT 插件扩展升级：改进排序与推荐方式，支持 MCP 事件自动化
IT之家 9 月 30 日消息，在北京时间 9 月 30 日凌晨 1 点举行的 OpenAI DevDay 2026 活动中，OpenAI 宣布升级插件扩展。 借助插件扩展，开发者可以直接在 ChatGPT 中构建丰富的应用内体验。现在， 用户可以为插件在侧边栏中设置专属入口 ，构建可供用户一边对话…
2
Multi-Modal Semantic Expansion with Constrained LLM Reranking for Conversational Music Recommendation
用多模态检索加约束重排搞定音乐对话推荐，参赛方案干货满满，值得算法人细读。
arXiv:2608.23484v1 Announce Type: new Abstract: We present Team Semiintelligencn's solution for the ACM RecSys 2026 TalkPlayData Challenge, addressing…
3
Show HN: From raw Point Cloud dataset to regular Grid index
点云数据转规则网格索引的开源工具，Squarenet 笛卡尔排序带来快速多维重排，为不规则 LiDAR 数据的卷积网络与 ANN 检索铺平预处理路径。
Interactive demo of a simple preprocessing tool for geometric datasets, all open source python package. Squarenet core sorting algorithm (Cartesian so…
4
The Parts Are Greater Than the Sum: Automated Task Sequencing for Efficient Training of Multi-Policy LLMs
多策略LLM训练新思路：用自动化任务排序取代单一LoRA，让各部件协同超越整体。
arXiv:2607.29601v1 Announce Type: new Abstract: Parameter-Efficient Fine-Tuning (PEFT) commonly adapts large language models using a single shared Low…
5
SciClaimSeekers at CheckThat! 2026: Retrieving Scientific Sources for Social Media Claims with LLM Reranking
用 LLM 重排序提升社交媒体言论的科学源检索精度，为事实核查提供高效方案。
arXiv:2607.24803v1 Announce Type: cross Abstract: Scientific claims often spread on social media faster than they can be verified, while posts rarely …
6
TabRank: Chain-of-Thought Distillation for Table Re-Rankers
用思维链蒸馏提升表格重排序效果，TabRank创新方法亮相。
arXiv:2607.25182v1 Announce Type: cross Abstract: The ability to retrieve relevant tables for answering questions is a key task for structured informa…
7
Post with links will no longer be nerfed on X
别错过这条重要更新：X平台终于对带链接的帖子不再降权，Hacker News 讨论贴附真实用户反馈
so nikita said recently in a reply to zuck's tweet that posting link in thread is no longer needed, the main post's links won't make your reach down. …
8
LLM-based Source Code Compression via Thresholded Symbol Ranking
用大语言模型结合阈值符号排序，打破通用压缩器局限，实现源代码高效压缩。
arXiv:2607.24192v1 Announce Type: cross Abstract: We study the problem of lossless compression of source code, motivated by the storage demands of lar…
9
I built a jav data API, need honest feedback
基于用户反馈难获取的痛点，这款Java数据API提供数据查询与集成能力，但需优化用户沟通机制
title says it all. honestly, i asked my users (50+) for feedback. 5-10 users used disposable emails, so they bounced. the rest just ghosted me. i even…
10
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…
11
When JSON Is Not Enough: Semantic Reliability of Schema-Constrained LLM Ordering Agents
当JSON不够用时，这项研究揭示了LLM排序代理在模式约束下的语义可靠性问题，值得关注。
arXiv:2607.18261v1 Announce Type: new Abstract: LLM agents are increasingly used as transaction compilers: a user states an intent in natural language…
12
PiCSAR: Probabilistic Confidence Selection And Ranking for Reasoning Chains
提出概率置信度选择与排序方法，优化大模型推理链，提升复杂推理的准确性与可解释性。
arXiv:2508.21787v3 Announce Type: replace-cross Abstract: Best-of-n sampling improves the accuracy of large language models (LLMs) and large reasoning…
13
Generative Pseudo-Labeling for Pre-Ranking with LLMs
提出一种利用大语言模型生成伪标签来优化预排序阶段的新方法，提升推荐系统效率。
arXiv:2602.20995v2 Announce Type: replace-cross Abstract: Pre-ranking is a critical stage in industrial recommendation systems, tasked with efficientl…
14
Internal Pluralism and the Limits of Pairwise Comparisons
当多元价值并存，成对比较为何失效？一篇挑战决策底层逻辑的硬核论文，值得深读。
arXiv:2607.02672v1 Announce Type: new Abstract: Local pairwise comparisons are a standard tool for learning how people want decision rules to work, e.…
15
AI Workflows Need Topological Sort
AI工作流依赖顺序执行，拓扑排序能避免依赖冲突，提升管线稳定性。
Article URL: https://arpitbhayani.me/blogs/ai-topological-sort/ Comments URL: https://news.ycombinator.com/item?id=48797087 Points: 1 # Comments: 0
16
Ask the Right Comparison:Bias-Aware Bayesian Active Top-$k$ Ranking with LLM Judges
LLM裁判偏爱长答案，传统偏置校正为何失效？贝叶斯主动Top-K排序新思路，直击评估偏差核心。
arXiv:2607.02104v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used as cheap, scalable judges that compare candidate ou…
17
Optimizing RAG Rerankers with LLM Feedback via Reinforcement Learning
利用强化学习与LLM反馈优化RAG重排序器，解决静态标注与生成脱节的关键难题。
arXiv:2604.02091v2 Announce Type: replace-cross Abstract: Rerankers play a pivotal role in refining retrieval results for Retrieval-Augmented Generati…
18
When Reranking Hurts: Uncertainty-Based Gating for Few-Shot Reranking
推翻重排序必有益的假设，提出无训练门控机制：只有不确定性高时再重排序，反而提升性能。
arXiv:2606.31087v1 Announce Type: cross Abstract: Few-shot selection typically assumes that reranking retrieved examples always improves performance. …
19
Can LLMs Rank? A Tale of Triads and Triage
一篇剖析大模型排序能力与偏差的前沿论文，揭示三元组与分诊机制背后的伦理陷阱，AI从业者必读。
arXiv:2606.30412v1 Announce Type: cross Abstract: From housing allocation for households experiencing homelessness to triage in emergency departments,…
20
Are LLMs Reliable Rankers? Rank Manipulation via Two-Stage Token Optimization
新论文揭示LLM排序漏洞：两阶段token优化可操控排名，可靠性存疑
arXiv:2510.06732v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used as rerankers in information retrieval, yet thei…