Abhishek Mishra
This stream will focus on dangerous capability evaluations for novel scientific risks across biology, chemistry, and materials science, alongside monitoring classifiers and scalable oversight for models working with automated laboratories.
Stream overview
Primary project descriptions:
- Dangerous capability evaluations for novel scientific risks, across biology, chemistry, and materials sciences, for models interacting with automated scientific labs.
- Building out inference-point / monitoring classifiers utilising LLMs orchestrating scientific models.
- Research on scientific-intelligence focused scalable oversight across training paradigms used which include automated labs as verifiers.
Mentorship style:
Standard (1-2 hours of weekly 1:1s)
Location during program:
SF Bay Area
London location preference:
Weak preference
Berkeley location preference:
Indifferent
Mentors
No items found.
Abhishek Mishra is Head of AI Safety at Lila Sciences. Mishra's research covers dangerous capability evaluations for novel scientific risks across biology, chemistry and materials science, and scalable oversight of models operating automated scientific labs.