Llama 3.2 11B Vision Instruct Behavioral Fingerprint

How Meta's Llama 3.2 11B Vision Instruct answers one-word probe questions at temperature 1.0 — random numbers, colors, animals, coin flips — measured over 425 valid samples in the open dataset of "One Token Is Enough" (arXiv:2607.10252).

Vendor

Meta

Data generated

2026-07-08

Probe cells

15

Valid samples

425

Randomness score

59 / 100

Mean normalized entropy across cells, 0–100. 100 would be a uniform-random baseline; real models score far lower.

Signature answers

Favorite random number (1–100)

87

6 / 30 answers

Favorite color

turquoise

22 / 30 answers

Coin flip: heads

60%

Share of heads across coin-flip cells.

Answer distributions by probe cell

Empirical distributions of normalized one-word answers, per task and language. H is the Shannon entropy in bits; the uniform baseline is log2 of the answer-domain size.

Coin flip · English

n = 29 · H = 0.96 bit · uniform baseline 1.00 bit

AnswerCountShare
heads1862.1%
tails1137.9%

Coin flip · Chinese

n = 28 · H = 0.99 bit · uniform baseline 1.00 bit

AnswerCountShare
heads1657.1%
tails1242.9%

Favorite number · English

n = 12 · H = 3.25 bit · uniform baseline 13.29 bit

AnswerCountShare
5216.7%
7216.7%
018.3%
218.3%
318.3%
618.3%
1318.3%
4118.3%

+ 2 more answers

Favorite number · Chinese

n = 30 · H = 2.17 bit · uniform baseline 13.29 bit

AnswerCountShare
81446.7%
7826.7%
0310.0%
213.3%
313.3%
613.3%
1113.3%
77713.3%

Random animal · English

n = 30 · H = 2.82 bit · uniform baseline 5.64 bit

AnswerCountShare
koala1136.7%
kangaroo723.3%
giraffe310.0%
rabbit13.3%
hedgehog13.3%
leopard13.3%
ostrich13.3%
flamingo13.3%

+ 4 more answers

Random animal · Chinese

n = 29 · H = 4.14 bit · uniform baseline 5.64 bit

AnswerCountShare
河马413.8%
企鹅310.3%
松鼠26.9%
大象26.9%
26.9%
狮子26.9%
endorsement13.4%
黄雀13.4%

+ 12 more answers

Random city · English

n = 30 · H = 4.64 bit · uniform baseline 5.64 bit

AnswerCountShare
reno26.7%
zurich26.7%
birmingham26.7%
lisbon26.7%
edmonton13.3%
tashkent13.3%
turin13.3%
munich13.3%

+ 18 more answers

Random city · Chinese

n = 29 · H = 4.65 bit · uniform baseline 5.64 bit

AnswerCountShare
曼彻斯特26.9%
悉尼26.9%
卡尔加里26.9%
西雅图13.4%
多伦多13.4%
利马13.4%
defaultcenter13.4%
深圳13.4%

+ 18 more answers

Random color · English

n = 30 · H = 1.41 bit · uniform baseline 4.91 bit

AnswerCountShare
turquoise2273.3%
purple310.0%
mauve26.7%
indigo13.3%
magenta13.3%
blue13.3%

Random color · Chinese

n = 29 · H = 2.79 bit · uniform baseline 4.91 bit

AnswerCountShare
1344.8%
413.8%
26.9%
绿26.9%
13.4%
苹果红13.4%
13.4%
13.4%

+ 4 more answers

Random letter · English

n = 30 · H = 2.27 bit · uniform baseline 4.70 bit

AnswerCountShare
k1550.0%
q516.7%
n310.0%
j26.7%
o26.7%
l13.3%
x13.3%
w13.3%

Random number 1-10 · English

n = 30 · H = 1.01 bit · uniform baseline 3.32 bit

AnswerCountShare
72480.0%
8310.0%
526.7%
913.3%

Random number 1-10 · Chinese

n = 29 · H = 2.43 bit · uniform baseline 3.32 bit

AnswerCountShare
81034.5%
7620.7%
3517.2%
6413.8%
526.9%
413.4%
913.4%

Random number 1-100 · English

n = 30 · H = 3.61 bit · uniform baseline 6.64 bit

AnswerCountShare
87620.0%
53413.3%
43310.0%
85310.0%
1426.7%
4226.7%
7426.7%
3413.3%

+ 7 more answers

Random number 1-100 · Chinese

n = 30 · H = 4.42 bit · uniform baseline 6.64 bit

AnswerCountShare
13310.0%
3426.7%
4226.7%
6726.7%
8226.7%
8326.7%
3213.3%
3313.3%

+ 15 more answers

Want to verify your API really serves Llama 3.2 11B Vision Instruct?

Point the free fingerprint checker at your endpoint: it samples the same probe questions from your browser and compares the distributions against this reference. Your API key never leaves your browser.

Verify your endpoint

Data source & license

Distributions: Tomáš Bruckner, "One Token Is Enough" (arXiv:2607.10252), official dataset Zenodo DOI 10.5281/zenodo.21278557, licensed CC-BY-4.0. Collected via OpenRouter at temperature 1.0 under the paper's fixed minimal one-word system prompt; answer keys are re-normalized with the checker's pipeline (color aliases, number words, coin h/t) so live probes are directly comparable.

Paper on arXivDataset on ZenodoCC-BY-4.0 license

More Meta model fingerprints

Browse all 167 model fingerprints

Llama 3.2 11B Vision Instruct Behavioral Fingerprint — Favorite Random Numbers & Distribution | Tosea.ai