1
Build an AI Error Explainer in Python
输入Python错误堆栈，AI自动分析原因并生成修复建议，告别手动排查
Stack traces are useful, but they are not always easy to act on quickly. When something breaks, you usually want more than the exception name. You wan…
2
IBM 股价大跌，CEO 克里希纳向投资者解释业绩未达预期原因
IT之家 7 月 15 日消息，7 月 14 日，IBM 的股价在该公司公布第二季度初步销售额低于预期后暴跌 25%， 创 1987 年以来最大单日跌幅 。 据 Business Insider 报道，IBM CEO 阿温德 · 克里希纳（Arvind Krishna）在周二致投资者的信中表示，公司…
3
Inside the Unfair Judge: A Mechanistic Interpretability Account of LLM-as-Judge Bias
用机械可解释性方法解剖LLM作为评判者时的内在偏见，揭示不公平机制根源
arXiv:2607.11871v1 Announce Type: cross Abstract: Existing studies of LLM-as-judge scoring bias work predominantly at the input-output level: they per…
4
Laguerre Geometry for Interpreting Large Language Models
Laguerre几何为理解大语言模型中的概念表示提供了精确的数学框架，将概念定义为区域而非单点或方向。
arXiv:2607.10578v1 Announce Type: new Abstract: Existing hypotheses represent a concept in an LLM as a single point, a linear direction, or a Gaussian…
5
Weight-Adjusted Gradients Reveal Parameter Importance and Failure Modes in LLMs
权重调整梯度的新方法揭示LLM参数重要性与失效模式，为模型可解释性提供新视角。
arXiv:2607.10803v1 Announce Type: cross Abstract: Understanding which parameters are influential in Large Language Models (LLMs) is central to improvi…
6
iLENS: Interpretable LLM-Guided Mixture-of-Experts for Neuroimaging Survival Analysis
将大语言模型与混合专家模型结合，攻克神经影像生存分析的可解释性难题，带来医学AI新思路。
arXiv:2607.08778v1 Announce Type: cross Abstract: Alzheimer's Disease (AD) is a complex neurodegenerative disorder that continues to impact millions o…
7
Mechanistic Interpretability of LLM Jailbreaks via Internal Attribution Graphs
用内部归因图揭示LLM越狱的运作机制，为AI安全提供新的可解释性视角。
arXiv:2607.07903v1 Announce Type: cross Abstract: Large language models (LLMs) exhibit remarkable capabilities but remain highly vulnerable to adversa…
8
Show HN: I made Claude explain like I'm 5, and AI fatigue disappeared
用“解释如五岁”规则让Claude变简单，AI疲劳瞬间消失，这招太聪明了！
I was burning out reading AI output. So I created the ELI5 Rule, ELI5 Rule is "explain-like-iam-five" It's simple. Try it. https://github.com/amebahea…
9
Anthropic found a hidden space where Claude puzzles over concepts
Anthropic用J-lens工具探入Claude Opus 4.6内部，发现了一个名为J-space的隐藏空间，揭示大模型处理概念的奇妙机制。
The AI firm Anthropic has developed a technique that has given it the clearest glimpse yet at what’s really going on inside large language models as t…
10
Before the Last Token: Diagnosing Final-Token Safety Probe Failures
揭秘大模型安全探针为何在最终token失效：早于最后一层的内部表征已暴露漏洞
arXiv:2605.12726v2 Announce Type: replace Abstract: Final-token safety probes monitor a single hidden state after prompt prefill, but jailbreak prompt…
11
Fair Document Valuation in LLM Summaries via Shapley Values
用Shapley值公平量化LLM摘要中各文档的贡献，解决价值分配不公问题，理论严谨。
arXiv:2505.23842v5 Announce Type: replace Abstract: Large Language Models (LLMs) increasingly power search engines and AI assistants that retrieve and…
12
Leveraging Color Naming for Image Enhancement
利用颜色命名规则，解决传统深度图像增强中可解释性差与参数化不足的两大痛点。
arXiv:2607.08185v1 Announce Type: new Abstract: Enhancing images to make them visually appealing is a persistent challenge in computer vision. Many de…
13
Large Language Models (LLMs) and Generative AI in Cybersecurity and Privacy: A Survey of Dual-Use Risks, AI-Generated Malware, Explainability, and Defensive Strategies
综述LLM和生成式AI在网络安全中的双刃剑效应，深度剖析AI生成恶意软件、可解释性及防御策略
arXiv:2607.06963v1 Announce Type: cross Abstract: Large Language Models (LLMs) and generative AI (GenAI) systems, such as ChatGPT, Claude, Gemini, LLa…
14
Towards Understanding Steering Strength
ICML 2026论文深入解析模型引导技术中的强度量化问题，为AI可解释性提供新视角。
arXiv:2602.02712v2 Announce Type: replace Abstract: A popular approach to post-training control of large language models (LLMs) is the steering of int…
15
Counterfactual Modeling with Fine-Tuned LLMs for Health Intervention Design and Sensor Data Augmentation
用微调大模型生成反事实解释，为健康干预设计和数据增强提供新思路，实用性强。
arXiv:2601.14590v3 Announce Type: replace Abstract: Counterfactual explanations (CFEs) provide human-centric interpretability by identifying the minim…
16
Dissociating the Internal Representations of Sycophancy in LLMs
从神经层面拆解大模型「拍马屁」行为的内部机制，一篇被ICML 2026研讨会接收的机械可解释性突破。
arXiv:2607.07003v1 Announce Type: new Abstract: Large Language Models (LLMs) frequently exhibit sycophancy, where they agree with a user's statement e…
17
A Good Initialization is All You Need for Faithful Visual Attribution
新研究发现良好初始化就能大幅提升视觉归因的忠实度，搜索式扰动方法成为关键。
arXiv:2607.06726v1 Announce Type: new Abstract: Faithful visual attribution identifies which image regions support a model prediction. Search-based pe…
18
CPU vs GPU: Why Large Language Models Need GPUs — What Really Happens After You Press Enter?
生动比喻揭秘为何GPU是大模型的大脑，从按下回车到万亿运算的探索之旅。
The moment you press Enter, billions of mathematical operations begin. Let's follow that journey. Every day, millions of people ask ChatGPT, Gemini, C…
19
Unsupervised Features Mining via Activation Geometry
利用激活几何进行无监督特征挖掘，无需人类标注即可揭示大模型内部表征
arXiv:2607.04222v1 Announce Type: new Abstract: Interpretability methods aim to reveal the features represented inside large language models (LLMs). M…
20
LLMs Encode Harmfulness and Refusal Separately
最新研究发现：LLM能分离编码“有害性”与“拒绝行为”，揭示安全机制新维度。
arXiv:2507.11878v5 Announce Type: replace Abstract: LLMs are trained to refuse harmful instructions, but do they truly understand harmfulness beyond j…