AI Desk · Model Identity · Evidence object

Tokenizer signature tk_066351de

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 Llama3 group.

Members
1
First observed
21 August 2026
Last observed
21 August 2026
Group
Llama3

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.

tk_6588ed5b · 6 modelstk_47615f2d · 4 models

The measured vector

Marginal prompt-token cost per string, as measured on Sao10K: Llama 3 8B Lunaris
All 50 rows

English and whitespace

en-prose13
en-long15
spaces-202
spaces-603
tabs-202
newlines-204
mixed-ws4

Digits

digits-92
digits-123
digits-3010
digits-sep6
float-long8

CJK

zh-common11
zh-long18
zh-rare18
ja-kana10
ja-kanji11
ko13

Other scripts

ru13
ar11
he23
hi14
th14
el12

Emoji

emoji-basic9
emoji-skin29
emoji-zwj-family15
emoji-zwj-x344
emoji-flags24
emoji-prof25

Rare Unicode

math30
boxdraw22
combining10
cjk-ext-b13
surrogates39
zalgo34
rtl-mix7

Code

py-code24
py-indent14
json19
html15
regex44
camel4
snake6

Repetition and encodings

rare-word-x529
repeat-tok20
base6434
hex11
url16
uuid26

Method: tokenizer fingerprinting, v1-50. Raw data: fingerprints.json · index.json · test-strings.json.