# Pre-training data detection on BookMIA (test)

Source: [Pre-training data detection on BookMIA (test)](<https://www.sota2.com/research/sota/pre-training-data-detection-on-bookmia-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-06T08:02:26.132Z

Evaluation of a model's ability to detect whether specific data from the BookMIA test set was used during pre-training, typically measured using AUROC and TPR@5%FPR.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Pre-training data detection |
| Dataset (including any stated split) | BookMIA (test) |
| Metric | AUROC |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 30 of 30 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [AUROC](<https://www.sota2.com/research/sota/pre-training-data-detection-on-bookmia-test.md?metric=AUROC>): Higher is better; unit: SCALAR.

- [TPR@5%FPR](<https://www.sota2.com/research/sota/pre-training-data-detection-on-bookmia-test.md?metric=TPR%405%25FPR>): Higher is better; unit: PERCENT.

## Recorded results

| Method | AUROC | Result date | Report paper |
| --- | --- | --- | --- |
| GDS | 0.99 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| GDS | 0.99 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| GDS | 0.99 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| GDS | 0.99 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| GDS | 0.99 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| FSD | 0.98 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| FSD | 0.98 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| FSD | 0.98 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| FSD | 0.98 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| FSD | 0.97 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k++ | 0.64 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k | 0.62 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k | 0.6 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k | 0.6 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k++ | 0.6 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| PPL | 0.59 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k++ | 0.58 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| PPL | 0.56 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k++ | 0.54 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| ZLib | 0.49 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k++ | 0.49 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| ZLib | 0.48 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k | 0.45 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| Min-k | 0.4 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| ZLib | 0.38 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| PPL | 0.29 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| PPL | 0.23 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| PPL | 0.22 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| ZLib | 0.2 | Not recorded | 69aa5f6c2d603e6de3e41799 |
| ZLib | 0.15 | Not recorded | 69aa5f6c2d603e6de3e41799 |

### GDS

AUROC: 0.99. Result ID: 69aa8a0f6dbb7075b95420da.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### GDS

AUROC: 0.99. Result ID: 69aa8a0f6dbb7075b95420db.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### GDS

AUROC: 0.99. Result ID: 69aa8a0f6dbb7075b95420dc.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

### GDS

AUROC: 0.99. Result ID: 69aa8a0f6dbb7075b95420dd.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### GDS

AUROC: 0.99. Result ID: 69aa8a0f6dbb7075b95420de.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### FSD

AUROC: 0.98. Result ID: 69aa8a0f6dbb7075b95420d5.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777c1e209131cea737dd.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### FSD

AUROC: 0.98. Result ID: 69aa8a0f6dbb7075b95420d7.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777c1e209131cea737dd.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

### FSD

AUROC: 0.98. Result ID: 69aa8a0f6dbb7075b95420d8.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777c1e209131cea737dd.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### FSD

AUROC: 0.98. Result ID: 69aa8a0f6dbb7075b95420d9.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777c1e209131cea737dd.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### FSD

AUROC: 0.97. Result ID: 69aa8a0f6dbb7075b95420d6.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa777c1e209131cea737dd.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### Min-k++

AUROC: 0.64. Result ID: 69aa8a0f6dbb7075b95420d1.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76f11e209131cea701dd.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### Min-k

AUROC: 0.62. Result ID: 69aa8a0f6dbb7075b95420ce.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea565b0.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### Min-k

AUROC: 0.6. Result ID: 69aa8a0f6dbb7075b95420cc.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea565b0.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### Min-k

AUROC: 0.6. Result ID: 69aa8a0f6dbb7075b95420cf.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea565b0.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### Min-k++

AUROC: 0.6. Result ID: 69aa8a0f6dbb7075b95420d3.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76f11e209131cea701dd.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### PPL

AUROC: 0.59. Result ID: 69aa8a0f6dbb7075b95420c4.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### Min-k++

AUROC: 0.58. Result ID: 69aa8a0f6dbb7075b95420d4.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76f11e209131cea701dd.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### PPL

AUROC: 0.56. Result ID: 69aa8a0f6dbb7075b95420c5.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### Min-k++

AUROC: 0.54. Result ID: 69aa8a0f6dbb7075b95420d0.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76f11e209131cea701dd.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### ZLib

AUROC: 0.49. Result ID: 69aa8a0f6dbb7075b95420c9.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"Pythia-6.9b"}. This record does not assert independent reproduction.

### Min-k++

AUROC: 0.49. Result ID: 69aa8a0f6dbb7075b95420d2.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76f11e209131cea701dd.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

### ZLib

AUROC: 0.48. Result ID: 69aa8a0f6dbb7075b95420ca.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"LLaMA-7B"}. This record does not assert independent reproduction.

### Min-k

AUROC: 0.45. Result ID: 69aa8a0f6dbb7075b95420cb.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea565b0.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### Min-k

AUROC: 0.4. Result ID: 69aa8a0f6dbb7075b95420cd.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72711e209131cea565b0.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

### ZLib

AUROC: 0.38. Result ID: 69aa8a0f6dbb7075b95420c7.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### PPL

AUROC: 0.29. Result ID: 69aa8a0f6dbb7075b95420c1.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### PPL

AUROC: 0.23. Result ID: 69aa8a0f6dbb7075b95420c3.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

### PPL

AUROC: 0.22. Result ID: 69aa8a0f6dbb7075b95420c2.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"GPT-J-6b"}. This record does not assert independent reproduction.

### ZLib

AUROC: 0.2. Result ID: 69aa8a0f6dbb7075b95420c6.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"Neo-2.7B"}. This record does not assert independent reproduction.

### ZLib

AUROC: 0.15. Result ID: 69aa8a0f6dbb7075b95420c8.

[Reporting paper](<https://www.sota2.com/research/paper/69aa5f6c2d603e6de3e41799.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a2b1e209131cea268e1.md>)

Recorded evaluation configuration: {"Target Model":"OPT-6.7b"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
