AI Desk · Model Identity · Evidence object
Tokenizer signature tk_59e2520c
An exact tokenizer signature: every model listed below returned the same marginal prompt-token count on all 50 test strings. Sharing a signature is strong evidence of a shared vocabulary. This signature sits in the Llama4 group.
- Members
- 2
- First observed
- 21 August 2026
- Last observed
- 21 August 2026
- Group
- Llama4
Models with this signature
Template variants of the same tokenizer
These signatures differ from this one by a uniform one-token shift on the affected strings: the same vocabulary behind a different chat template.
The measured vector
Marginal prompt-token cost per string, as measured on Meta: Llama 4 MaverickAll 50 rows
English and whitespace
en-prose14
en-long17
spaces-203
spaces-603
tabs-203
newlines-204
mixed-ws6
Digits
digits-93
digits-124
digits-3010
digits-sep7
float-long9
CJK
zh-common11
zh-long12
zh-rare18
ja-kana9
ja-kanji12
ko10
Other scripts
ru10
ar11
he10
hi12
th8
el14
Emoji
emoji-basic10
emoji-skin20
emoji-zwj-family11
emoji-zwj-x333
emoji-flags16
emoji-prof19
Rare Unicode
math24
boxdraw21
combining10
cjk-ext-b16
surrogates37
zalgo31
rtl-mix6
Code
py-code25
py-indent15
json20
html16
regex42
camel6
snake6
Repetition and encodings
rare-word-x530
repeat-tok40
base6435
hex11
url17
uuid27
Method: tokenizer fingerprinting, v1-50. Raw data: fingerprints.json · index.json · test-strings.json.