David Vos @davidvos.bsky.social · 27/02/2025As it turns out: It's one of the longest tokens without special characters. When considering special characters, the top ones look like this. 110
David Vos @davidvos.bsky.social · 27/02/2025Today's funny tokenization realization: The GPT4 tokenizer has 1 single token for the entire lower- and uppercase alphabet. 210
David Vos @davidvos.bsky.social · 13/12/2024Congratulations dr. Zihan Wang! It was an honor to be your paranymph. 090
David Vos @davidvos.bsky.social · 18/11/2024Furthermore, among other things, I really enjoy teaching 👨🏻🏫, Scotland 🏴, going for long (slow but hopefully increasingly fast) runs 🏃, reading fiction 📖, and connecting with new people :) 110
David Vos @davidvos.bsky.social · 18/11/2024I love thinking about how to get the right information to the right people in the right way. This quest started about two years ago, when I worked together with @mersault.bsky.social on using LLMs for data management. The picture is me presenting @trl-research.bsky.social. 131
David Vos @davidvos.bsky.social · 09/11/2024‘Based on what you know about me, draw a picture of what you think my current life looks like.’ A @bsky.app introduction by ChatGPT. 020