AdaBPE adapts BPE merge thresholds for mixed language and LaTeX
On five LaTeX and natural-language datasets, the method improves encoding and domain-specific model performance without hurting general language.
2 stories, newest first · one card per story
On five LaTeX and natural-language datasets, the method improves encoding and domain-specific model performance without hurting general language.
Simon Willison covers a LaTeX renderer inspired by Tiny Little Office's URL documents and agent llms.txt.
Today 1.5K readers · 2.4K views · 757 reads · all time 189K views