Kamal Ndousse
Anthropic researcher; previously OpenAI; multi-agent RL and language model training.
- Role
- researcher
- Currently at
- Anthropic
- twitter.com/kamalndousse
- GitHub
- github.com/kandouss
- Scholar
- scholar.google.com/citations
- Papers
- 5
Cite
Notes
Only stored in your browser.
Authored papers
5Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
arXiv 2024
Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
preprint
Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
arXiv 2022
Discovering Language Model Behaviors with Model-Written Evaluations
arXiv 2022
Constitutional AI: Harmlessness from AI Feedback
arXiv 2022
Affiliations
Previously
Frequent co-authors
10from 5 papers
Amanda Askell
researcher
Deep Ganguli
researcher
Jared Kaplan
co-founder / Chief Science Officer
Nova DasSarma
researcher
Shauna Kravec
researcher
Yuntao Bai
researcher
Andy Jones
researcher
Anna Chen
researcher
Ben Mann
founder
Catherine Olsson
researcher