Berkeley Existential Risk Initiative — Language Model Alignment Research

Open Philanthropy recommended a grant of $30,000 over three years to the Berkeley Existential Risk Initiative to support the Algorithmic Alignment Group (AAG). Led by Dylan Hadfield-Menell, AAG...

Open Philanthropy recommended a grant of $500,000 over two years to support the Berkeley Existential Risk Initiative’s collaboration with the Center for Human-Compatible (CHAI) to work on research relevant...

Open Philanthropy recommended a grant of $45,000 to the Berkeley Existential Risk Initiative to support a project led by Stephen Casper on latent adversarial training methods for large language models....