P
R
Panas
Research
We are an independent AI safety research lab.
We develop novel evaluation methods drawn from frontier AI safety research, and turn them into tools that stress-test AI systems at scale.
Services
1. Automated Red Teaming
2. Mechanistic Interpretability (In development, Partial)
3. Post-safety training (Coming Soon)
1. Automated Red Teaming
2. Mechanistic Interpretability (In development, Partial)
3. Post-safety training (Coming Soon)