AI Desk · Model Identity · Evidence object

Tokenizer signature tk_6588ed5b

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: 6
First observed: 21 August 2026
Last observed: 21 August 2026
Group: Llama3

Models with this signature

Meta: Llama 3.1 70B Instructmeta-llama/llama-3.1-70b-instructMeta Meta: Llama 3.1 8B Instructmeta-llama/llama-3.1-8b-instructMeta Meta: Llama 3.2 3B Instructmeta-llama/llama-3.2-3b-instructMeta Meta: Llama 3.3 70B Instructmeta-llama/llama-3.3-70b-instructMeta Sao10K: Llama 3.1 Euryale 70B v2.2sao10k/l3.1-euryale-70bSao10k Sao10K: Llama 3.3 Euryale 70Bsao10k/l3.3-euryale-70bSao10k

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_47615f2d · 4 models tk_066351de · 1 model

The measured vector

Marginal prompt-token cost per string, as measured on Meta: Llama 3.1 70B Instruct

All 50 rows

English and whitespace

en-prose14

en-long16

spaces-203

spaces-603

tabs-203

newlines-204

mixed-ws5

Digits

digits-93

digits-124

digits-3010

digits-sep7

float-long9

CJK

zh-common12

zh-long18

zh-rare18

ja-kana11

ja-kanji12

ko14

Other scripts

ru14

ar12

he24

hi15

th14

el13

Emoji

emoji-basic10

emoji-skin30

emoji-zwj-family15

emoji-zwj-x345

emoji-flags24

emoji-prof26

Rare Unicode

math31

boxdraw23

combining11

cjk-ext-b13

surrogates40

zalgo35

rtl-mix8

Code

py-code25

py-indent15

json20

html16

regex44

camel5

snake6

Repetition and encodings

rare-word-x530

repeat-tok21

base6435

hex11

url17

uuid27

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