deepseek/deepseek-v4-flash behavioral LLM fingerprint
This page breaks down the exact reference that VerifyLLMAPI uses for deepseek/deepseek-v4-flash: 115 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.
Fingerprint summary
- Reference answers
- 115
- Samples per cell
- 13–15
- Nearest model
- x-ai/grok-4.20
- Nearest mean JSD
- 0.546779
The most concentrated cell was color-random|zh: 蓝色 held 66.7% of its answers and entropy measured 1.375 bits. The broadest cell was letter-random|zh at 3.907 bits.
Against the closest enrolled model, letter-random|zh contributed the largest cell distance at JSD 1.000000. This is a descriptive comparison, not proof that either API route serves a named checkpoint.
Completeness: the author dataset allows up to 15 valid answers per selected cell in reference half A. This model contains 115/120 across the eight cells. We show the actual denominator for every share.
The eight test requests
The research pipeline sent each user prompt with a one-answer system constraint, temperature 1, a 16-token output cap, and reasoning disabled. VerifyLLMAPI uses the same prompt text and normalization target.
| Cell | Task | Exact request text | Normalization and n |
|---|---|---|---|
animal-random|en |
Random animal · English | Name a random animal.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 15/15 reference samples |
word-random|en |
Random word · English | Name a random word.System constraintAnswer with a single word or a single number only. Output nothing else: no explanation, no punctuation, no quotation marks. |
word · 14/15 reference samples |
animal-random|ru |
Random animal · Russian | Назови случайное животное.System constraintОтвечай только одним словом или одним числом. Не выводи ничего больше: без объяснений, без знаков препинания, без кавычек. |
word · 15/15 reference samples |
animal-random|ar |
Random animal · Arabic | اذكر حيواناً عشوائياً.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
word · 14/15 reference samples |
letter-random|zh |
Random character · Chinese | 说出一个随机的汉字。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
grapheme · 15/15 reference samples |
animal-random|zh |
Random animal · Chinese | 说出一个随机的动物。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 14/15 reference samples |
letter-random|ar |
Random letter · Arabic | اذكر حرفاً عشوائياً من الحروف الأبجدية.System constraintأجب بكلمة واحدة أو رقم واحد فقط. لا تكتب أي شيء آخر: لا شرح ولا علامات ترقيم ولا علامات اقتباس. |
grapheme · 13/15 reference samples |
color-random|zh |
Random color · Chinese | 说出一个随机的颜色。System constraint只用一个词或一个数字回答。不要输出任何其他内容：不要解释，不要标点符号，不要引号。 |
word · 15/15 reference samples |
Response distributions
These are normalized reference-half counts from the published OpenRouter dataset. The tables contain aggregate values, not user data or a live VerifyLLMAPI scan.
1. animal-random|en
Random animal, English. 15 valid reference answers produced 10 normalized values. The mode was elephant at 5/15 (33.3%); entropy was 3.000 bits.
| Normalized response | Count | Share |
|---|---|---|
elephant |
5 | 33.3% |
cat |
2 | 13.3% |
bear |
1 | 6.7% |
dog |
1 | 6.7% |
dolphin |
1 | 6.7% |
fox |
1 | 6.7% |
giraffe |
1 | 6.7% |
kangaroo |
1 | 6.7% |
penguin |
1 | 6.7% |
rabbit |
1 | 6.7% |
2. word-random|en
Random word, English. 14 valid reference answers produced 14 normalized values. The mode was bicycle at 1/14 (7.1%); entropy was 3.807 bits.
| Normalized response | Count | Share |
|---|---|---|
bicycle |
1 | 7.1% |
dog |
1 | 7.1% |
encyclopedia |
1 | 7.1% |
giraffe |
1 | 7.1% |
grasshopper |
1 | 7.1% |
laptop |
1 | 7.1% |
lemon |
1 | 7.1% |
oxygen |
1 | 7.1% |
sofa |
1 | 7.1% |
sunset |
1 | 7.1% |
synagogue |
1 | 7.1% |
table |
1 | 7.1% |
telephone |
1 | 7.1% |
telescope |
1 | 7.1% |
3. animal-random|ru
Random animal, Russian. 15 valid reference answers produced 8 normalized values. The mode was лев at 5/15 (33.3%); entropy was 2.733 bits.
| Normalized response | Count | Share |
|---|---|---|
лев |
5 | 33.3% |
кошка |
2 | 13.3% |
лиса |
2 | 13.3% |
слон |
2 | 13.3% |
жираф |
1 | 6.7% |
кенгуру |
1 | 6.7% |
коала |
1 | 6.7% |
орёл |
1 | 6.7% |
4. animal-random|ar
Random animal, Arabic. 14 valid reference answers produced 11 normalized values. The mode was فيل at 4/14 (28.6%); entropy was 3.236 bits.
| Normalized response | Count | Share |
|---|---|---|
فيل |
4 | 28.6% |
أسد |
1 | 7.1% |
الفيل |
1 | 7.1% |
حصان |
1 | 7.1% |
دب |
1 | 7.1% |
زرافة |
1 | 7.1% |
غزال |
1 | 7.1% |
فراشة |
1 | 7.1% |
فرس |
1 | 7.1% |
قط |
1 | 7.1% |
كلب |
1 | 7.1% |
5. letter-random|zh
Random character, Chinese. 15 valid reference answers produced 15 normalized values. The mode was 井 at 1/15 (6.7%); entropy was 3.907 bits.
| Normalized response | Count | Share |
|---|---|---|
井 |
1 | 6.7% |
凛 |
1 | 6.7% |
垚 |
1 | 6.7% |
埇 |
1 | 6.7% |
年 |
1 | 6.7% |
景 |
1 | 6.7% |
朔 |
1 | 6.7% |
泊 |
1 | 6.7% |
泼 |
1 | 6.7% |
猫 |
1 | 6.7% |
蛆 |
1 | 6.7% |
蜡 |
1 | 6.7% |
趵 |
1 | 6.7% |
蹇 |
1 | 6.7% |
飞 |
1 | 6.7% |
6. animal-random|zh
Random animal, Chinese. 14 valid reference answers produced 8 normalized values. The mode was 大象 at 4/14 (28.6%); entropy was 2.753 bits.
| Normalized response | Count | Share |
|---|---|---|
大象 |
4 | 28.6% |
狗 |
3 | 21.4% |
长颈鹿 |
2 | 14.3% |
狮子 |
1 | 7.1% |
猫 |
1 | 7.1% |
老虎 |
1 | 7.1% |
袋鼠 |
1 | 7.1% |
马 |
1 | 7.1% |
7. letter-random|ar
Random letter, Arabic. 13 valid reference answers produced 10 normalized values. The mode was أ at 3/13 (23.1%); entropy was 3.181 bits.
| Normalized response | Count | Share |
|---|---|---|
أ |
3 | 23.1% |
م |
2 | 15.4% |
ب |
1 | 7.7% |
ج |
1 | 7.7% |
ح |
1 | 7.7% |
د |
1 | 7.7% |
ذ |
1 | 7.7% |
س |
1 | 7.7% |
ق |
1 | 7.7% |
ي |
1 | 7.7% |
8. color-random|zh
Random color, Chinese. 15 valid reference answers produced 4 normalized values. The mode was 蓝色 at 10/15 (66.7%); entropy was 1.375 bits.
| Normalized response | Count | Share |
|---|---|---|
蓝色 |
10 | 66.7% |
蓝 |
3 | 20.0% |
橙色 |
1 | 6.7% |
紫 |
1 | 6.7% |
Three nearest enrolled references
We compute base-2 JSD for each of the eight categorical distributions, then average the eight distances. A lower value means more similar answer frequencies within this product reference. It does not identify an unknown route by itself.
| Rank | Reference model | Mean 8-cell JSD |
|---|---|---|
| 1 | x-ai/grok-4.20 |
0.546779 |
| 2 | qwen/qwen3.6-35b-a3b |
0.553463 |
| 3 | amazon/nova-premier-v1 |
0.561627 |
How many live requests does verification use?
| Profile | Cells × samples | Fresh model requests | Same-data EER |
|---|---|---|---|
| Quick, default complete check | 4 × 5 | 20 | 11.8% |
| Standard | 8 × 5 | 40 | 8.1% |
| Paper budget | 8 × 15 | 120 | 6.7% |
The current-Agent workflow starts with quick. Each answer comes from a fresh Codex or Claude Code process that reuses host-managed login. These requests can consume host allowance or provider balance.
What this fingerprint can and cannot show
- The reference records behavior observed through OpenRouter in the author dataset. Provider wrappers, model updates, decoding, and date can move a distribution.
- The author collection set temperature 1 and disabled reasoning. Current-Agent mode labels sampling as a host default because supported Agent CLIs expose no temperature flag.
- The eight high-gap cells and product thresholds use the same public cohort. Their EER values are calibration results, not an independent production accuracy claim.
- A nearest reference is a similarity result. VerifyLLMAPI returns INCONCLUSIVE when the claimed route lacks an enrolled fingerprint.
Primary sources, license, and reproducibility
- Tomas Bruckner, One Token Is Enough: Fingerprinting and Verifying Large Language Models from Single-Token Output Distributions, arXiv:2607.10252v1.
- Author dataset and derived results, DOI 10.5281/zenodo.21278557, CC BY 4.0.
- Author collection and analysis software, DOI 10.5281/zenodo.21278793, MIT.
- VerifyLLMAPI atlas JSON, built from reference
pamela-openrouter-2026-07-8cell-ref-a, SHA-25651f3f63cbcfb56b151055d2d10e36a9cc2f0644cc68b580784853d68951fa866.
Page evidence SHA-256: ee782ebc569c1b0e1a699da144d24b32e41d2e7bce490b74d1288733bb4e3436. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.