Transluce

Infrastructure for understanding AI

Transluce is a non-profit research lab building the public tech stack for scalable oversight of AI

Our Work

WeirdChat

Research

WeirdChat

A catalog of unexpected AI behaviors, discovered automatically

Toward A Public Science of Model Behavior

Essay

Toward A Public Science of Model Behavior

Our vision for an open scientific ecosystem for measuring model behaviors

Predictive Concept Decoders

Research

Predictive Concept Decoders

Training scalable end-to-end interpretability assistants

Latest News

→
ResearchJuly 21, 2026

WeirdChat

A catalog of unexpected AI behaviors, discovered automatically

EssayJuly 9, 2026

Toward A Public Science of Model Behavior

Our vision for an open scientific ecosystem for measuring model behaviors

DemoJune 17, 2026

Introducing Analysis Plans

A framework for verifiable analysis of AI behavior

DemoFebruary 11, 2026

Diagnosing a performance regression on Terminal-Bench with Docent

Why does GPT-5.1 Codex underperform GPT-5 Codex?