Preference Model
San Francisco, United States
◆ About Preference Model
A superintelligence data research company building reinforcement learning environments that resist reward hacks.
◆ What sets it apart
Its reinforcement-learning environments and open-source Karotte framework are engineered to withstand reward hacking through hardened sandboxes, graders, and evaluation infrastructure.
◆ Sectors
◆ Business
ModelSaaS
CustomerFrontier AI laboratories and organizations developing advanced AI models
◆ Notable
- Raised $16 million in seed funding led by Andreessen Horowitz
- Built reinforcement-learning environments for several frontier AI labs
- Released Karotte, an open-source framework for building reward-hack-resistant RL environments
◆ Funding rounds
◆ Explore the market
◆ FAQ
What does Preference Model do?
A superintelligence data research company building reinforcement learning environments that resist reward hacks.
Where is Preference Model based?
Preference Model is based in San Francisco, United States.
How much has Preference Model raised?
Preference Model has raised $16M in tracked funding and is currently at the seed stage.
Who has invested in Preference Model?
Investors in Preference Model include Andreessen Horowitz, SignalFire, South Park Commons, Scale Angels, Fei-Fei Li, Ian Goodfellow and Julian Schrittwieser.