dmodel

d_model builds RL environments where agents learn to do interpretability and alignment research, and runs similar agents in its own research.

dmodel.ai · updated Sep 29, 2026

Trust & data

How the vendor protects buyer material, where its data comes from, and where it can run.

Data sources: Synthetic / generated
Source:dmodel.ai
Quote
Synthetic / generated: Using this technique, we can generate large numbers of programs in an unsupervised manner, and then label them fully automatically to get many prompts for training our probe.
Not publicly listed: SOC 2, Trust center, Other certifications, Deployment, Exclusivity, Stated confidentiality practices

Offerings: RL environments
Source:dmodel.ai
Domains: Science Coding
Sources:dmodel.ai (1)dmodel.ai/nullability
Public samples & benchmarks: NullabilityEval
Steering Characters with Interpretability notebook
Sources:dmodel.ai/nullability dmodel.ai/steering_characters
Not publicly listed: Delivery formats, Named customers

Every change to these facts is kept in the public claims log. Record co_7241642471ce.