Othello GPT Insights
Martin Wattenberg discusses the world model aspect of Othello GPT and the causal experiments conducted to understand its internal model. He emphasizes the importance of skepticism and understanding the nuances of neural networks' memorization and generalization capabilities.In this clip
From this podcast
The Gradient
Martin Wattenberg: ML Visualization and Interpretability
Related Questions