Tokenization in DeepSeek R1

Channel:

LLMs Explained - Aggregate Intellect - AI.SCIENCE

Subscribers:

22,300

Published on March 17, 2025 5:00:48 AM ● Video Link: https://www.youtube.com/watch?v=tUpErbDUQes

Duration: 0:00

164 views

1

In this video, we dissect token boundary bias, a common pitfall in models like Stability AI’s tokenizer, where unexpected spaces or punctuation splits derail outputs and how DeepSeek fixed this in their R1 model.


Where else to find us:
https://www.linkedin.com/in/amirfzpr/
https://aisc.substack.com/
   / @ai-science  
https://lu.ma/aisc-llm-school
https://maven.com/aggregate-intellect/