Skip to main content
Back to catalog

Mistral

3 glitch tokens

These tokens have been publicly documented on this model or model family. Inclusion does not mean that every version still shows the same behavior.

\uEFC0

A Unicode Private Use Area character that became a single Mistral-family token; it heads the verified under-trained list for nearly every Mistral-tokenizer model.
MistralUnder-trained

ICENSE

The word “LICENSE” minus its first letter — an artifact of open-source license text flooding the corpus.
MistralUnspeakable

NdEx

A legal-text fragment of the same family as GPT-NeoX's ` NdEx`, but from the Mistral tokenizer (no leading space).
MistralUnspeakable