Exploring the new multimodal model from AI2.| seantrott.substack.com
How do VLMs combine their modalities?| seantrott.substack.com
Modern language models predict "tokens", not words—but what exactly are tokens?| seantrott.substack.com
