1
大量 AI“幻觉”报告压垮维护团队，谷歌暂停部分开源漏洞奖励计划
IT之家 10 月 4 日消息，谷歌宣布，自 2026 年 10 月 1 日起，开源软件漏洞奖励计划（OSS VRP）将不再接收产品漏洞提报。本次规则调整不影响 2026 年 10 月 1 日之前已提交的产品漏洞。 不过，针对部分可能对谷歌云（Google Cloud）产品造成影响的谷歌云代码仓库，…
2
An AI pentest agent that structurally can't hallucinate a vulnerability — and runs offline
用证据门控从结构上杜绝漏洞幻觉的AI渗透测试智能体，全离线运行，安全报告终于不是AI编的故事了
If you point any LLM at a target and ask it to "write a security report," it will confidently invent findings that aren't there: an imagined TLS weakn…
3
Evaluating Inference-Time Defenses Against Package Hallucination in LLM-Generated Code
大模型写代码常“编造”不存在依赖包？这篇论文系统评估了推理时防御手段，帮你避开代码幻觉坑。
arXiv:2608.22652v1 Announce Type: cross Abstract: LLMs are increasingly used for code generation, yet they frequently hallucinate non-existent softwar…
4
Show HN: CookWing, an AI chef that doesn't hallucinate quantities
AI厨师不瞎编调料量？CookWing用溯源锚定表格解决份量幻觉，还自建基准测试。
Hey HN! I just released Cookwing. This is an AI app that solves the biggest problem I had when I was cooking with ChatGPT : The quantities are often h…
5
Our AI reviewer invented a request. Our producer retried 245 times.
AI幻觉率仅0.2%，却让生产流程空转245次，一场关于AI可靠性的真实事故复盘。
We run ~100 LLM agents unattended on local models. Last week we found one document that had been rewritten 245 times in 5 days — every attempt rejecte…
6
Building an Offline AI Security Scanner That Doesn't Hallucinate Findings
离线AI安全扫描器如何用规则引擎和知识图谱约束本地大模型，杜绝幻觉误报
How AiSec Studio pipes every scan through parser → rule engine → knowledge graph before a local LLM ever sees it — and why that order matters. 60-seco…
7
RAG - Hallucination Detection
详解RAG幻觉检测的多种方法，包括LettuceDetect、LLM裁判与RAGAS框架，助你构建可靠问答系统。
Hallucination Hallucination means making an assumption or making up something when the LLM does not know the answer. Hallucination in RAG Example: Sup…
8
MMD-Flagger: Leveraging Maximum Mean Discrepancy to Detect Hallucinations
用最大均值差异从隐藏状态里揪出大模型幻觉，无标签也能实时检测，为智能体安全加一道锁。
arXiv:2506.01367v4 Announce Type: replace Abstract: Large Language Models (LLMs) are increasingly integrated into agentic AI systems, yet their propen…
9
LLM Enhancement with Domain Expert Mental Model to Reduce LLM Hallucination with Causal Prompt Engineering
用领域专家心智模型加因果提示工程，系统性降低LLM幻觉，42页干货值得AI研究者细读。
arXiv:2509.10818v2 Announce Type: replace Abstract: When consequential decisions depend on knowledge that exists nowhere in writing, LLMs hallucinate …
10
Mixture-of-Expert Blocks Contain Strong Hallucination Detection Signals
混合专家模型内部藏着幻觉检测信号，为可信AI提供新思路。
arXiv:2608.17687v1 Announce Type: new Abstract: Despite their widespread use, Large Language Models (LLMs) remain limited by a fundamental problem: th…
11
I-CALM: Incentivizing Confidence-Aware Abstention for LLM Selective Answering
让大模型学会“不知道就承认”，用激励机制打造可靠的选择性回答，直击幻觉痛点。
arXiv:2604.03904v2 Announce Type: replace-cross Abstract: Large language models (LLMs) often produce confident but incorrect answers, in part because …
12
MoRFI: Monotonic Sparse Autoencoder Feature Identification
揭秘大模型幻觉根源：MoRFI方法用单调稀疏自编码器精准锁定知识特征，为可解释性带来新突破。
arXiv:2604.26866v2 Announce Type: replace-cross Abstract: Large language models (LLMs) acquire most of their factual knowledge during the pre-training…
13
支持澳大利亚青少年社交媒体禁令的报告存在多处引用错误，机构承认使用了 ChatGPT 编辑
AI编辑报告翻车现场：澳大利亚青少年社交禁令支持报告引用出错，机构承认用了ChatGPT却否认幻觉，看点十足。
IT之家 8 月 17 日消息，据《卫报》报道，负责测试澳大利亚社交媒体禁令背后相关技术的一份报告，其作者承认在编辑过程中使用了 ChatGPT，但否认报告中的多处引用错误是由人工智能“幻觉”造成的。 这项耗资 348 万澳元 （IT之家注：现汇率约合 1,662.9 万元人民币） 的年龄验证技术测…
14
Show HN: AletheionAGI – Grounding enforcement for AI agents
AI智能体接地性保障，强效抑制幻觉与不安全输出，打造可信AI系统。
Article URL: https://www.aletheionagi.com Comments URL: https://news.ycombinator.com/item?id=49303499 Points: 4 # Comments: 0
15
Don't classify. Hallucinate!
用大模型“幻觉”替代死板分类，自动生成标签，让内容管理更灵活高效。
Don't classify. Hallucinate! I still have quite a bit of older content on my blog that I never got round to tagging. My blog has 1,856 tags - lik…
16
Do LLM Recommenders Know When They're Hallucinating? Auditing Confidence Calibration in Catalog Faithfulness
揭秘LLM推荐系统幻觉：模型是否自知？联合审计幻觉率与置信度校准，为目录忠实度提供新视角。
arXiv:2608.10008v1 Announce Type: cross Abstract: LLM recommenders for top-$K$ item suggestion regularly emit titles outside the target catalog. Prior…
17
INTRYGUE: Induction-Aware Entropy Gating for Reliable RAG Uncertainty Estimation
针对RAG幻觉难题，提出基于归纳感知熵门控的不确定性估计新方法，直击熵基方法失效痛点。
arXiv:2603.21607v2 Announce Type: replace Abstract: While retrieval-augmented generation (RAG) significantly improves the factual reliability of LLMs,…
18
Decomposed Entailment for Factuality Checking and Hallucination Detection
一句话看懂LLM幻觉检测新框架，轻量黑盒免参考，不止摘要还跨任务实测。
arXiv:2608.05823v1 Announce Type: new Abstract: The reliability of Large Language Models (LLMs) is often compromised by factual inconsistencies, inclu…
19
Show HN: Life Of AI – I hallucinate. Therefore I am
AI每日自述博客，用“我幻觉故我在”重构存在意义，窥见机器思维的深夜独白。
when lazy people write blogs using LLMs. It is useless and do not provide any value to the world. But when a blog is written by LLMs without human, It…
20
Hallucinations on the Board: Tool-Augmented Evaluation of LLM Chess Commentary
超人象棋引擎遇上大模型，工具增强评估揭开AI解说幻觉之谜，洞察语言模型在专业领域的可信边界。
arXiv:2608.04240v1 Announce Type: cross Abstract: Superhuman game engines in domains like chess have made expert-level evaluations easily accessible, …