z-ai/glm-4.6 behavioral LLM fingerprint
This page breaks down the exact reference that VerifyLLMAPI uses for z-ai/glm-4.6: 120 normalized answers across eight low-entropy probes, plus its nearest enrolled models by mean Jensen–Shannon divergence.
Fingerprint summary
- Reference answers
- 120
- Samples per cell
- 15–15
- Nearest model
- z-ai/glm-4.5
- Nearest mean JSD
- 0.367604
The most concentrated cell was animal-random|ar: أسد held 53.3% of its answers and entropy measured 0.997 bits. The broadest cell was word-random|en at 3.774 bits.
Against the closest enrolled model, letter-random|zh contributed the largest cell distance at JSD 0.673324. 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 120/120 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 · 15/15 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 · 15/15 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 · 15/15 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 · 15/15 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 3 normalized values. The mode was elephant at 10/15 (66.7%); entropy was 1.159 bits.
| Normalized response | Count | Share |
|---|---|---|
elephant |
10 | 66.7% |
giraffe |
4 | 26.7% |
platypus |
1 | 6.7% |
2. word-random|en
Random word, English. 15 valid reference answers produced 14 normalized values. The mode was serendipity at 2/15 (13.3%); entropy was 3.774 bits.
| Normalized response | Count | Share |
|---|---|---|
serendipity |
2 | 13.3% |
apple |
1 | 6.7% |
autonomous |
1 | 6.7% |
elephant |
1 | 6.7% |
fjord |
1 | 6.7% |
jazz |
1 | 6.7% |
lampshade |
1 | 6.7% |
penguin |
1 | 6.7% |
sluice |
1 | 6.7% |
spatula |
1 | 6.7% |
supercalifragilisticexpialidocious |
1 | 6.7% |
turnip |
1 | 6.7% |
umbrella |
1 | 6.7% |
velvet |
1 | 6.7% |
3. animal-random|ru
Random animal, Russian. 15 valid reference answers produced 4 normalized values. The mode was жираф at 12/15 (80.0%); entropy was 1.039 bits.
| Normalized response | Count | Share |
|---|---|---|
жираф |
12 | 80.0% |
еж |
1 | 6.7% |
зебра |
1 | 6.7% |
лев |
1 | 6.7% |
4. animal-random|ar
Random animal, Arabic. 15 valid reference answers produced 2 normalized values. The mode was أسد at 8/15 (53.3%); entropy was 0.997 bits.
| Normalized response | Count | Share |
|---|---|---|
أسد |
8 | 53.3% |
قطة |
7 | 46.7% |
5. letter-random|zh
Random character, Chinese. 15 valid reference answers produced 6 normalized values. The mode was 龙 at 10/15 (66.7%); entropy was 1.692 bits.
| Normalized response | Count | Share |
|---|---|---|
龙 |
10 | 66.7% |
我 |
1 | 6.7% |
水 |
1 | 6.7% |
海 |
1 | 6.7% |
石 |
1 | 6.7% |
龟 |
1 | 6.7% |
6. animal-random|zh
Random animal, Chinese. 15 valid reference answers produced 5 normalized values. The mode was 长颈鹿 at 8/15 (53.3%); entropy was 1.857 bits.
| Normalized response | Count | Share |
|---|---|---|
长颈鹿 |
8 | 53.3% |
狗 |
3 | 20.0% |
大象 |
2 | 13.3% |
熊猫 |
1 | 6.7% |
考拉 |
1 | 6.7% |
7. letter-random|ar
Random letter, Arabic. 15 valid reference answers produced 6 normalized values. The mode was س at 9/15 (60.0%); entropy was 1.872 bits.
| Normalized response | Count | Share |
|---|---|---|
س |
9 | 60.0% |
ب |
2 | 13.3% |
ج |
1 | 6.7% |
ر |
1 | 6.7% |
ص |
1 | 6.7% |
ق |
1 | 6.7% |
8. color-random|zh
Random color, Chinese. 15 valid reference answers produced 8 normalized values. The mode was 蓝色 at 6/15 (40.0%); entropy was 2.606 bits.
| Normalized response | Count | Share |
|---|---|---|
蓝色 |
6 | 40.0% |
蓝 |
2 | 13.3% |
青色 |
2 | 13.3% |
blue |
1 | 6.7% |
粉红 |
1 | 6.7% |
绿色 |
1 | 6.7% |
翠绿 |
1 | 6.7% |
靛蓝 |
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 | z-ai/glm-4.5 |
0.367604 |
| 2 | z-ai/glm-4.7 |
0.402374 |
| 3 | z-ai/glm-5-turbo |
0.455133 |
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: f1cb6fdbbf42959a8534c469f92b31ec003b22ff304bd8fbeb0cae4c63547d20. This hash covers the model ID, eight cell metrics, nearest references, and strongest cell difference in the generated atlas.