Eliciting Latent Knowledge: Prizewinning Submission
The Alignment Research Center recently offered prizes for proposals for Eliciting Latent Knowledge(ELK). The challenge in ELK is to devise a strategy for tra...
Recent Posts
Training Transformers using Reinforcement Learning
The capabilities of Large Language Models have increased dramatically in the past few years, particularly with GPT-3, and they are frankly amazing. Watching ...
2021 MineRL BASALT Competition (2nd Place)
In the fall of 2021, I participated in the 2021 MineRL BASALT competition (Submission Code). This competition, organized by the Center for Human-Compatible A...