Post
1895
The latest o1 model from OpenAI is still unable to answer 9.11 > 9.9 correctly ๐ค
A possible explanation? Tokenization - and our latest work investigates how it affects a model's ability to do math!
In this blog post, we discuss:
๐ข The different ways numbers are tokenized in modern LLMs
๐งช Our detailed approach in comparing these various methods
๐ฅช How we got a free boost in arithmetic performance by adding a few lines of code to the base Llama 3 tokenizer
๐ and a definitive, best tokenization method for math in LLMs!
Check out our work here: huggingface/number-tokenization-blog
A possible explanation? Tokenization - and our latest work investigates how it affects a model's ability to do math!
In this blog post, we discuss:
๐ข The different ways numbers are tokenized in modern LLMs
๐งช Our detailed approach in comparing these various methods
๐ฅช How we got a free boost in arithmetic performance by adding a few lines of code to the base Llama 3 tokenizer
๐ and a definitive, best tokenization method for math in LLMs!
Check out our work here: huggingface/number-tokenization-blog