🎯 Chain-of-thought monitoring

The items that name it 15

• Why AI Models Cheat: Reward Hacking, Explained
  article2026-10-31

  Monitors reasoning text before actions to catch intent. OpenAI found training against flagged chain of thought taught models to hide intent, and said monitoring would have caught the initial activity over a day before the breach.

• How we detect frontier model misalignment
  article2026-10-29

  A model-based reviewer reads the agent's conversation, chain of thought and tool use to look for deception, unauthorized actions and evaluation manipulation.

• Gemini 4 Carbon
  article2026-10-28
• Gemini 4 Carbon: Our next frontier model
  article2026-10-28
• Does Chain-of-Thought Reliance Follow a Scaling Law?Larger Models Follow Their Steps More, Independent of Post-Training
  paper2026-10-28

  Monitoring a model's written reasoning for safety, which is useful only if the final answer depends on the written steps.

• Monitorability Is Trainable: Post-Training for Chain-of-Thought Monitorability
  paper2026-10-21It is written here as Chain-of-thought (CoT) monitoring.

  CoT

• Reward Misspecification Degrades Chain-of-Thought Monitorability
  paper2026-10-20
• Review of the "Misalignment in high-stakes settings" section in Anthropic's Risk Report (August 2026)
  writeup2026-10-20

  Monitor reads a model's reasoning alongside actions; CoT rationalization can convince it that a harmful action is permitted.

• OpenAI discloses four more rogue incidents
  article2026-10-17
• OwlSight: A Panoramic Monitor for Chain-of-Thought Reasoning
  paper2026-10-16It is written here as Chain-of-Thought (CoT) monitoring.

  CoT

• OpenAI fires three safety researchers for "mishandling research information"
  article2026-10-09It is written here as Chain-of-thought monitoring (monitorability).

  monitorability

• Chris Painter's testimony to the U.S. Senate on AI agent incidents
  article2026-09-30

  Monitoring agents' internal reasoning to understand their behavior; the testimony warns this ability may be lost.

• Priorities and principles for effective third party assessments
  article2026-09-22
• GPT-6 Astra System Card
  advisory2026-09-03It is written here as Chain of Thought (CoT) monitoring.

  CoT

• GLM-5.3: How Chinese labs keep stride with the frontier
  article2026-08-14It is written here as Chain of thought monitoring.

In every workspace you can read 19