Shouldn't all caps normalised to tokens like low caps? There are no separate tok... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

miohtama 8 months ago | parent | context | favorite | on: Nano Banana can be prompt engineered for nuanced A...

Shouldn't all caps normalised to tokens like low caps? There are no separate tokens for all caps and low caps in Llama, or at least not in the past.

minimaxir 8 months ago [–]

Looking at the tokenizer for the older Llama 2 model, the tokenizer has capital letters in it: https://huggingface.co/meta-llama/Llama-2-7b-hf

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact