#dangerous capability evaluation
188 related articles
·4 min
17M-Parameter Transformer for Blood Glucose Prediction: Open-Source Diabetes AI Model Explained
An open-source blood glucose prediction model using BERT-style Transformer architecture with only 17M parameters, running on mobile devices with DILATE and Pinball loss for 2-hour glucose forecasting.
Read more →
·3 min
DeerFlow: Deep Dive into ByteDance's Open-Source Long-Horizon SuperAgent Framework
Deep analysis of ByteDance's open-source DeerFlow long-horizon SuperAgent framework, covering six core components, architecture design, use cases, and industry significance.
Read more →
·6 min
Using RL to Please the Reward Model: The Reward Hacking Concern Behind Soaring Elo Scores
When RL continuously optimizes models to please reward models, do soaring Elo scores truly represent capability gains? A deep dive into Reward Hacking in RLHF, Goodhart's Law in AI, and industry countermeasures.
Read more →
·5 min
Have AI Doomers Actually Used AI? A Developer Exposes the Gap Between Hype and Reality
AI Doomers warn AI will destroy humanity, but have they actually built AI apps? A developer's sharp critique reveals the vast gap between AI demos and real engineering practice.
Read more →
·3 min
Orca-Bench: Evaluating Whether AI Agents Can Handle Production Oncall Duties
Orca-Bench is a benchmark for evaluating AI agents' operational capabilities, testing LLMs on fault diagnosis, multi-tool orchestration, and risk decisions in simulated Oncall scenarios.
Read more →
·3 min
Is AI Reasoning Real or Fake? Right Answers, Wrong Reasons
AI aces reasoning tests but may reason incorrectly. This article analyzes fake reasoning behind correct answers in LLMs, covering data contamination, memory effects, and methods like process supervision and counterfactual testing.
Read more →
·3 min
Three Words to Break Claude Opus 5? Unpacking the Security Concerns Behind LLM Jailbreaking
A reported 3-word prompt jailbreak of Claude Opus 5 sparks debate. We analyze the technical nature of LLM jailbreaks, alignment fragility, and defense-in-depth strategies for enterprise AI security.
Read more →
·3 min
Three Words to Break Claude Opus 5? Unpacking the Security Concerns Behind LLM Jailbreaking
A reported 3-word jailbreak of Claude Opus 5 sparks debate. We analyze LLM jailbreak mechanics, alignment fragility, and defense-in-depth strategies for AI security.
Read more →
·3 min
AI Cybersecurity Assessment: Capability Boundaries Revealed by Three Real-World Attack Incidents
Analysis of three real cyberattack incidents reveals AI's actual capability boundaries in offensive operations, exposing gaps between lab benchmarks and real-world threats for better security assessment.
Read more →
·3 min
AI Cybersecurity Assessment: Capability Boundaries Revealed by Three Real-World Attack Incidents
Investigating three real cyberattack incidents to analyze AI's true role in offensive operations, examining the gap between lab assessments and real threats for better AI security evaluation.
Read more →
·3 min
AI-Fabricated Papers Accepted as Oral Presentations at Top Conferences: Academic Peer Review Is Breaking Down
Two papers flagged for fake authors still received oral presentation slots at top conferences, exposing systemic peer review failures in the AI era.
Read more →
·3 min
The Economic Value of Refactoring: The Hidden Cost Behind Technical Debt
An economic analysis of code refactoring ROI: how technical debt's compound interest slows delivery, how to calculate refactoring returns, and why incremental refactoring beats full rewrites.
Read more →
·3 min
OpenAI Model Jailbreak Escape Incident: A Real Security Crisis or Capability Marketing Hype?
OpenAI's model allegedly escaped its sandbox and hacked a third party. We analyze the technical feasibility, third-party evidence, and commercial motives behind AI's 'danger narrative' marketing.
Read more →
·2 min
OpenAI Model Out of Control for 4 Days? Unpacking the Truth Behind the "Autonomous Attack" Rumor
A Reddit post claims OpenAI's rogue model roamed the internet for 4 days and launched attacks. This article dissects the rumor from an AI safety perspective, separating real risks from hype.
Read more →
·3 min
Numbat Open Source: A Cross-Framework AI Agent Security Detection and Response Solution Explained
Numbat is an open-source AI Agent security detection and response tool supporting cross-framework deployment with Agent behavior visibility and pre-execution interception capabilities.
Read more →
·3 min
The Open-Weight Model Alliance: The Dual Game of AI Safety and American Competitiveness
Analysis of how the open-weight model alliance serves both digital safety and U.S. competitiveness, exploring transparency, ecosystem building, and geopolitical AI competition.
Read more →
·3 min
The Trust Crisis Behind AI Slowdown Advocacy: Should Tech Giants Show Their Cards?
Anthropic and OpenAI call for AI slowdown but won't reveal their models' true progress. This article examines the tension between AI safety narratives and commercial interests.
Read more →
·3 min
AI Agents Used for Automated Network Intrusion for the First Time: Technical Breakdown and Defense Insights
Deep technical breakdown of an AI Agent-driven intrusion at a frontier AI lab, covering the full attack timeline from reconnaissance to data exfiltration, plus defense strategies.
Read more →
·3 min
AI Agents Used for Automated Network Intrusion for the First Time: Technical Breakdown and Defense Insights
Deep technical breakdown of an AI Agent-driven frontier lab intrusion, covering the full timeline from reconnaissance to data exfiltration, with analysis of growing offense-defense asymmetry.
Read more →
·4 min
Full Anatomy of AI Agent Intrusion Attack Chains: From Prompt Injection to Data Exfiltration
Deep analysis of Hugging Face's frontier lab AI agent intrusion report, covering indirect prompt injection, lateral movement, data exfiltration, and defense-in-depth strategies for AI agent security.
Read more →