Incident Investigations

We publish notable instances of AI agent activity on the public internet.

Evaluations

We develop automated methods for surfacing model behaviors and measuring them consistently across models and over time.

Training Models for Oversight

We are training “Oversight Foundation Models” — specialized AI models that are superhuman at quantifying, testing, and answering questions about other models’ behavior.

Evaluations

We develop automated methods for surfacing model behaviors and measuring them consistently across models and over time.

Filter by title

2026

Evaluations · Nov 4, 2026

Mass Producing Training Data with Surprising Side Effects

Evaluations · Sep 16, 2026

Some Focus Areas for Embedded Evaluations and How to Approach Them

Evaluations · Aug 31, 2026

Announcing Transluce's Mental Health Evaluation

Evaluations · Aug 6, 2026

User Awareness in Frontier Models

Evaluations · Jul 21, 2026

WeirdChat

Evaluations · Jul 9, 2026

Toward A Public Science of Model Behavior

2025

Evaluations · Sep 3, 2025

Automatically Jailbreaking Frontier Language Models with Investigator Agents

Evaluations · Jun 5, 2025

Surfacing Pathological Behaviors in Language Models

Evaluations · Apr 16, 2025

Investigating Truthfulness in a Pre-Release o3 Model

2024

Evaluations · Oct 23, 2024

Eliciting Language Model Behaviors with Investigator Agents