Datasetpublic

G-reen/fastdetector-test-stat-test

Auto-Generated FastDetector Dataset Dataset: G-reen/fastdetector-test-stat-test Globals Config: config/globals_test.toml Analysis Config: config/analysis.toml Rows: 13,622 Evaluation Results Prompt Subsets: 4 (direct_reference, indirect_reference, revise, rewrite) Generator Configs: 10 (claude-opus-4-5-20251101 (Temp: Unknown), claude-opus-5 (Temp: Unknown), gemini-3.1-pro-preview (Temp: Unknown), gemini-3.8-flash (Temp: Unknown), gpt-5.4 (Temp: Unknown)… See the full description on the dataset page: https://huggingface.co/datasets/G-reen/fastdetector-test-stat-test.

sourceHugging Faceupdated 2h agoView on Hugging Face

0likes615downloads

Dataset Card

Auto-Generated FastDetector Dataset

—Dataset: G-reen/fastdetector-test-stat-test
—Globals Config: config/globals_test.toml
—Analysis Config: config/analysis.toml
—Rows: 13,622

Evaluation Results

—Prompt Subsets: 4 (directreference, indirectreference, revise, rewrite)
—Generator Configs: 10 (claude-opus-4-5-20251101 (Temp: Unknown), claude-opus-5 (Temp: Unknown), gemini-3.1-pro-preview (Temp: Unknown), gemini-3.8-flash (Temp: Unknown), gpt-5.4 (Temp: Unknown), gpt-5.6-sol (Temp: Unknown), grok-4.3 (Temp: Unknown), kimi-k3 (Temp: Unknown), minimax-m3 (Temp: Unknown), qwen3.7-max (Temp: Unknown))
—Classifiers: 19 (EditLens Roberta-Large Score, EditLens Roberta-Large Bucket, EditLens Llama-3.2-3B Score, EditLens Llama-3.2-3B Bucket, EditLens Giga RoBERTa-Large Score, EditLens Giga RoBERTa-Large Bucket, EditLens Giga Llama-3.2-3B Score, EditLens Giga Llama-3.2-3B Bucket, Perplexity (Llama-3.2-3B-Instruct), Perplexity (Llama-3.2-3B), Entropy (Llama-3.2-3B-Instruct), Entropy (Llama-3.2-3B), Top-p Outliers (Llama-3.2-3B-Instruct), Top-p Outliers (Llama-3.2-3B), Top-k Outliers (Llama-3.2-3B-Instruct), Top-k Outliers (Llama-3.2-3B), FastDetectGPT (Llama-3.2-3B-Instruct), FastDetectGPT (Llama-3.2-3B), Binoculars)
—Filter Conditions: cosdist >= 0.03 OR softngram >= 0.06
—Evaluation / Validation Rows: 12,259 / 1,363 (validation_size = 0.1)
—Base Columns: original (Human), final_response (AI)

The best classifier was EditLens Giga Llama-3.2-3B Score with an AUROC of 0.9442. The hardest prompt subset was rewrite with a TPR of 0.0962, and the hardest generator config was claude-opus-5 (Temp: Unknown) with a TPR of 0.1190.

Classifier	Threshold	AUROC	TPR	FPR	Accuracy	F1
✔️ EditLens Giga Llama-3.2-3B Score	0.1014	0.9442	0.6413	0.0033	0.8190	0.7799
EditLens Llama-3.2-3B Score	0.2389	0.9214	0.6163	0.0057	0.8053	0.7599
EditLens Giga RoBERTa-Large Score	0.1910	0.9107	0.5930	0.0041	0.7944	0.7426
EditLens Roberta-Large Score	0.4747	0.8930	0.3411	0.0031	0.6690	0.5076
EditLens Llama-3.2-3B Bucket	0.0000	0.8135	0.6297	0.0037	0.8130	0.7710
EditLens Giga RoBERTa-Large Bucket	0.0000	0.7925	0.5881	0.0043	0.7919	0.7386
EditLens Roberta-Large Bucket	0.0000	0.7832	0.5945	0.0370	0.7788	0.7288
EditLens Giga Llama-3.2-3B Bucket	0.0000	0.7770	0.5544	0.0004	0.7770	0.7131
Binoculars	0.9673	0.6201	0.0448	0.0033	0.5208	0.0855
Top-p Outliers (Llama-3.2-3B)	0.0201	0.5801	0.0400	0.0049	0.5175	0.0765
Entropy (Llama-3.2-3B-Instruct)	1.4637	0.5397	0.0188	0.0090	0.5049	0.0365
Perplexity (Llama-3.2-3B-Instruct)	4.7220	0.5270	0.0115	0.0060	0.5028	0.0226
Top-k Outliers (Llama-3.2-3B-Instruct)	0.0295	0.5132	0.0221	0.0080	0.5071	0.0429
Top-p Outliers (Llama-3.2-3B-Instruct)	0.0312	0.5009	0.0197	0.0077	0.5060	0.0384
Perplexity (Llama-3.2-3B)	3.9421	0.4860	0.0096	0.0100	0.4998	0.0189
FastDetectGPT (Llama-3.2-3B-Instruct)	3.2271	0.4790	0.0067	0.0039	0.5014	0.0132
Top-k Outliers (Llama-3.2-3B)	0.0183	0.4788	0.0109	0.0073	0.5018	0.0215
Entropy (Llama-3.2-3B)	1.4228	0.4720	0.0093	0.0114	0.4989	0.0182
❗ FastDetectGPT (Llama-3.2-3B)	-2.8000	0.4520	0.0235	0.0079	0.5078	0.0456

Classifier metrics averaged within each prompt and generator subset:

Subset	Average AUROC	Average TPR	Average FPR	Average Accuracy	Average F1
✔️ Model: claude-opus-4-5-20251101 (Temp: Unknown)	0.6982	0.2979	0.0082	0.6449	0.3643
Prompt: revise	0.6926	0.3212	0.0081	0.6565	0.3702
Model: minimax-m3 (Temp: Unknown)	0.6906	0.2667	0.0072	0.6298	0.3412
Model: kimi-k3 (Temp: Unknown)	0.6801	0.2808	0.0073	0.6367	0.3540
Prompt: indirect_reference	0.6647	0.2265	0.0066	0.6100	0.3057
Model: gpt-5.6-sol (Temp: Unknown)	0.6600	0.2037	0.0074	0.5982	0.2783
Prompt: direct_reference	0.6595	0.2850	0.0079	0.6385	0.3545
Model: qwen3.7-max (Temp: Unknown)	0.6560	0.3065	0.0081	0.6492	0.3644
Model: gemini-3.8-flash (Temp: Unknown)	0.6551	0.3126	0.0068	0.6529	0.3716
Model: gpt-5.4 (Temp: Unknown)	0.6499	0.1818	0.0082	0.5868	0.2566
Model: gemini-3.1-pro-preview (Temp: Unknown)	0.6447	0.2638	0.0070	0.6284	0.3312
Model: grok-4.3 (Temp: Unknown)	0.6395	0.2763	0.0069	0.6347	0.3443
Model: claude-opus-5 (Temp: Unknown)	0.5963	0.1190	0.0071	0.5560	0.1856
❗ Prompt: rewrite	0.5719	0.0962	0.0067	0.5448	0.1583

✔️ marks the best AUROC, ❗ the worst.

Statistics of Interest

[image] [image] [image] [image]

Appendix

Table of contents 1. Univariate Analysis 2. Correlation Heatmap 3. Distance Histograms 4. Distance Histograms per Prompt Subset 5. Distance Histograms per Generator Config Subset 6. Classifier: EditLens Roberta-Large Score - Performance: - Thresholding: - Classification Histograms: 7. Classifier: EditLens Roberta-Large Bucket - Performance: - Thresholding: - Classification Histograms: 8. Classifier: EditLens Llama-3.2-3B Score - Performance: - Thresholding: - Classification Histograms: 9. Classifier: EditLens Llama-3.2-3B Bucket - Performance: - Thresholding: - Classification Histograms: 10. Classifier: EditLens Giga RoBERTa-Large Score - Performance: - Thresholding: - Classification Histograms: 11. Classifier: EditLens Giga RoBERTa-Large Bucket - Performance: - Thresholding: - Classification Histograms: 12. Classifier: EditLens Giga Llama-3.2-3B Score - Performance: - Thresholding: - Classification Histograms: 13. Classifier: EditLens Giga Llama-3.2-3B Bucket - Performance: - Thresholding: - Classification Histograms: 14. Classifier: Perplexity (Llama-3.2-3B-Instruct) - Performance: - Thresholding: - Classification Histograms: 15. Classifier: Perplexity (Llama-3.2-3B) - Performance: - Thresholding: - Classification Histograms: 16. Classifier: Entropy (Llama-3.2-3B-Instruct) - Performance: - Thresholding: - Classification Histograms: 17. Classifier: Entropy (Llama-3.2-3B) - Performance: - Thresholding: - Classification Histograms: 18. Classifier: Top-p Outliers (Llama-3.2-3B-Instruct) - Performance: - Thresholding: - Classification Histograms: 19. Classifier: Top-p Outliers (Llama-3.2-3B) - Performance: - Thresholding: - Classification Histograms: 20. Classifier: Top-k Outliers (Llama-3.2-3B-Instruct) - Performance: - Thresholding: - Classification Histograms: 21. Classifier: Top-k Outliers (Llama-3.2-3B) - Performance: - Thresholding: - Classification Histograms: 22. Classifier: FastDetectGPT (Llama-3.2-3B-Instruct) - Performance: - Thresholding: - Classification Histograms: 23. Classifier: FastDetectGPT (Llama-3.2-3B) - Performance: - Thresholding: - Classification Histograms: 24. Classifier: Binoculars - Performance: - Thresholding: - Classification Histograms:

Univariate Analysis

Every statistic the report does arithmetic on, over the 12,259-row evaluation split. Invalid counts rows whose value is missing or non-finite; those rows are excluded from the other columns.

Statistic	N	Mean	Median	Std	Min	Max
jaccard_1	12,259	0.6431	0.7214	0.2317	0.0000	1.0000
jaccard_2	12,259	0.7701	0.8807	0.2459	0.0210	1.0000
levenshtein	12,259	2396.6086	1768.0000	2580.4291	35.0000	79984.0000
softngram	12,259	0.6053	0.6833	0.3236	0.0000	1.0000
cosdist	12,259	0.1819	0.1104	0.1830	-0.0061	0.9648
bertscore	12,259	0.1333	0.1400	0.0657	0.0029	0.7966
bertscore_precision	12,259	0.1321	0.1389	0.0640	0.0016	0.7497
bertscore_recall	12,259	0.1335	0.1394	0.0726	0.0009	0.8287
moverscore	12,259	0.5332	0.5763	0.1664	0.0848	1.3348
reranker	12,259	-5.5643	-7.0625	4.2713	-10.9375	17.1250
EditLens Roberta-Large Score (Human)	12,259	0.0460	0.0204	0.0712	0.0064	0.9838
EditLens Roberta-Large Score (AI)	12,259	0.4073	0.3141	0.3418	0.0064	0.9996
EditLens Roberta-Large Bucket (Human)	12,259	0.0407	0.0000	0.2243	0.0000	3.0000
EditLens Roberta-Large Bucket (AI)	12,259	1.1242	1.0000	1.2018	0.0000	3.0000
EditLens Llama-3.2-3B Score (Human)	12,259	0.0300	0.0159	0.0405	0.0003	0.4905
EditLens Llama-3.2-3B Score (AI)	12,259	0.3888	0.3434	0.3077	0.0004	1.0000
EditLens Llama-3.2-3B Bucket (Human)	12,259	0.0037	0.0000	0.0605	0.0000	1.0000
EditLens Llama-3.2-3B Bucket (AI)	12,259	1.0624	1.0000	1.0805	0.0000	3.0000
EditLens Giga RoBERTa-Large Score (Human)	12,259	0.0085	0.0039	0.0253	0.0013	0.6720
EditLens Giga RoBERTa-Large Score (AI)	12,259	0.4379	0.3448	0.4006	0.0014	1.0000
EditLens Giga RoBERTa-Large Bucket (Human)	12,259	0.0055	0.0000	0.0941	0.0000	3.0000
EditLens Giga RoBERTa-Large Bucket (AI)	12,259	1.2825	1.0000	1.2837	0.0000	3.0000
EditLens Giga Llama-3.2-3B Score (Human)	12,259	0.0050	0.0027	0.0114	0.0005	0.1835
EditLens Giga Llama-3.2-3B Score (AI)	12,259	0.3764	0.2672	0.3664	0.0006	1.0000
EditLens Giga Llama-3.2-3B Bucket (Human)	12,259	0.0004	0.0000	0.0202	0.0000	1.0000
EditLens Giga Llama-3.2-3B Bucket (AI)	12,259	1.0892	1.0000	1.2076	0.0000	3.0000
Perplexity (Llama-3.2-3B-Instruct) (Human)	12,259	15.7390	14.0895	7.0593	2.9954	65.3062
Perplexity (Llama-3.2-3B-Instruct) (AI)	12,259	15.6716	13.7293	11.8022	1.0071	537.9031
Perplexity (Llama-3.2-3B) (Human)	12,259	11.9412	10.9053	4.9894	1.2068	47.0632
Perplexity (Llama-3.2-3B) (AI)	12,259	12.6835	11.2033	9.6719	1.0034	612.3403
Entropy (Llama-3.2-3B-Instruct) (Human)	12,259	2.5079	2.4940	0.4341	1.0731	4.2297
Entropy (Llama-3.2-3B-Instruct) (AI)	12,259	2.4476	2.4369	0.4937	0.0365	7.1878
Entropy (Llama-3.2-3B) (Human)	12,259	2.3870	2.3884	0.3813	0.2294	3.7963
Entropy (Llama-3.2-3B) (AI)	12,259	2.4282	2.4268	0.3990	0.0100	5.5116
Top-p Outliers (Llama-3.2-3B-Instruct) (Human)	12,259	0.0552	0.0539	0.0120	0.0215	0.1231
Top-p Outliers (Llama-3.2-3B-Instruct) (AI)	12,259	0.0555	0.0540	0.0144	0.0000	0.2000
Top-p Outliers (Llama-3.2-3B) (Human)	12,259	0.0422	0.0418	0.0094	0.0064	0.0796
Top-p Outliers (Llama-3.2-3B) (AI)	12,259	0.0396	0.0390	0.0120	0.0000	0.1818
Top-k Outliers (Llama-3.2-3B-Instruct) (Human)	12,259	0.0967	0.0922	0.0348	0.0064	0.2630
Top-k Outliers (Llama-3.2-3B-Instruct) (AI)	12,259	0.0962	0.0918	0.0415	0.0001	0.5000
Top-k Outliers (Llama-3.2-3B) (Human)	12,259	0.0784	0.0743	0.0308	0.0039	0.2192
Top-k Outliers (Llama-3.2-3B) (AI)	12,259	0.0824	0.0773	0.0380	0.0001	0.5000
FastDetectGPT (Llama-3.2-3B-Instruct) (Human)	12,259	-1.8409	-1.7987	1.6619	-10.5708	5.9897
FastDetectGPT (Llama-3.2-3B-Instruct) (AI)	12,259	-1.9475	-1.9468	1.9309	-14.4647	56.7824
FastDetectGPT (Llama-3.2-3B) (Human)	12,259	-0.1500	-0.1090	1.0304	-5.8795	2.7665
FastDetectGPT (Llama-3.2-3B) (AI)	12,259	0.1333	0.0632	1.5958	-10.4853	13.9080
Binoculars (Human)	12,259	0.8468	0.8545	0.0696	0.0379	1.0063
Binoculars (AI)	12,259	0.8720	0.8717	0.0616	0.0355	1.2666

Correlation Heatmap

Pearson correlation between every statistic of interest, computed over the rows where both statistics are present.

[image]

Distance Histograms

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Distance Histograms per Prompt Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Distance Histograms per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Roberta-Large Score

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.8930	0.3411	0.0031	0.6690	0.5076
Prompt: direct_reference	7,190	0.8926	0.4423	0.0033	0.7195	0.6119
Prompt: indirect_reference	6,488	0.8992	0.3055	0.0028	0.6514	0.4670
✔️ Prompt: revise	7,114	0.9682	0.4150	0.0037	0.7057	0.5850
❗ Prompt: rewrite	3,726	0.7432	0.0671	0.0021	0.5325	0.1255
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.9212	0.4551	0.0040	0.7256	0.6238
Model: claude-opus-5 (Temp: Unknown)	2,440	0.8221	0.0877	0.0025	0.5426	0.1609
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.8992	0.3645	0.0040	0.6803	0.5327
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.9385	0.5098	0.0023	0.7537	0.6743
Model: gpt-5.4 (Temp: Unknown)	2,404	0.8517	0.1839	0.0042	0.5899	0.3095
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.8890	0.1951	0.0024	0.5963	0.3258
Model: grok-4.3 (Temp: Unknown)	2,268	0.8902	0.4065	0.0035	0.7015	0.5766
Model: kimi-k3 (Temp: Unknown)	2,360	0.9101	0.3695	0.0025	0.6835	0.5386
Model: minimax-m3 (Temp: Unknown)	2,376	0.8879	0.3367	0.0017	0.6675	0.5031
Model: qwen3.7-max (Temp: Unknown)	2,562	0.9187	0.4902	0.0039	0.7432	0.6562

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.4747.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Roberta-Large Bucket

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.7832	0.5945	0.0370	0.7788	0.7288
Prompt: direct_reference	7,190	0.7903	0.6022	0.0337	0.7843	0.7363
Prompt: indirect_reference	6,488	0.7682	0.5703	0.0428	0.7637	0.7071
✔️ Prompt: revise	7,114	0.8858	0.7990	0.0357	0.8816	0.8710
❗ Prompt: rewrite	3,726	0.5987	0.2313	0.0354	0.5980	0.3653
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.8287	0.6823	0.0373	0.8225	0.7935
Model: claude-opus-5 (Temp: Unknown)	2,440	0.6239	0.2820	0.0361	0.6230	0.4279
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.8134	0.6585	0.0412	0.8086	0.7748
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.8755	0.7729	0.0345	0.8692	0.8553
Model: gpt-5.4 (Temp: Unknown)	2,404	0.6823	0.3977	0.0374	0.6801	0.5542
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.7393	0.5080	0.0334	0.7373	0.6591
Model: grok-4.3 (Temp: Unknown)	2,268	0.8138	0.6552	0.0397	0.8078	0.7732
Model: kimi-k3 (Temp: Unknown)	2,360	0.8163	0.6602	0.0364	0.8119	0.7782
Model: minimax-m3 (Temp: Unknown)	2,376	0.7828	0.5909	0.0337	0.7786	0.7275
Model: qwen3.7-max (Temp: Unknown)	2,562	0.8490	0.7237	0.0398	0.8419	0.8207

Thresholding:

—Direction: higher_is_ai
—Swept for f1 with a found threshold of 0.0000.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Llama-3.2-3B Score

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.9214	0.6163	0.0057	0.8053	0.7599
Prompt: direct_reference	7,190	0.9332	0.6885	0.0083	0.8401	0.8115
Prompt: indirect_reference	6,488	0.9089	0.5465	0.0043	0.7711	0.7048
✔️ Prompt: revise	7,114	0.9788	0.8088	0.0051	0.9019	0.8918
❗ Prompt: rewrite	3,726	0.8142	0.2308	0.0043	0.6133	0.3738
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.9349	0.7006	0.0071	0.8467	0.8205
Model: claude-opus-5 (Temp: Unknown)	2,440	0.8320	0.3156	0.0057	0.6549	0.4777
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.9312	0.6395	0.0055	0.8170	0.7775
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.9560	0.7424	0.0055	0.8684	0.8495
Model: gpt-5.4 (Temp: Unknown)	2,404	0.9024	0.5042	0.0050	0.7496	0.6681
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.9343	0.5605	0.0048	0.7779	0.7162
Model: grok-4.3 (Temp: Unknown)	2,268	0.9265	0.6605	0.0053	0.8276	0.7930
Model: kimi-k3 (Temp: Unknown)	2,360	0.9288	0.6636	0.0059	0.8288	0.7949
Model: minimax-m3 (Temp: Unknown)	2,376	0.9192	0.6389	0.0059	0.8165	0.7769
Model: qwen3.7-max (Temp: Unknown)	2,562	0.9464	0.7276	0.0062	0.8607	0.8393

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.2389.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Llama-3.2-3B Bucket

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.8135	0.6297	0.0037	0.8130	0.7710
Prompt: direct_reference	7,190	0.8429	0.6882	0.0039	0.8421	0.8134
Prompt: indirect_reference	6,488	0.7824	0.5675	0.0034	0.7821	0.7225
✔️ Prompt: revise	7,114	0.9082	0.8198	0.0045	0.9076	0.8988
❗ Prompt: rewrite	3,726	0.6299	0.2619	0.0021	0.6299	0.4144
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.8563	0.7156	0.0048	0.8554	0.8319
Model: claude-opus-5 (Temp: Unknown)	2,440	0.6509	0.3049	0.0033	0.6508	0.4662
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.8290	0.6593	0.0016	0.8288	0.7939
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.8788	0.7604	0.0039	0.8782	0.8620
Model: gpt-5.4 (Temp: Unknown)	2,404	0.7558	0.5150	0.0042	0.7554	0.6780
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.7869	0.5772	0.0040	0.7866	0.7301
Model: grok-4.3 (Temp: Unknown)	2,268	0.8383	0.6790	0.0035	0.8377	0.8071
Model: kimi-k3 (Temp: Unknown)	2,360	0.8377	0.6780	0.0034	0.8373	0.8065
Model: minimax-m3 (Temp: Unknown)	2,376	0.8191	0.6406	0.0034	0.8186	0.7793
Model: qwen3.7-max (Temp: Unknown)	2,562	0.8764	0.7557	0.0047	0.8755	0.8585

Thresholding:

—Direction: higher_is_ai
—Swept for f1 with a found threshold of 0.0000.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Giga RoBERTa-Large Score

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.9107	0.5930	0.0041	0.7944	0.7426
Prompt: direct_reference	7,190	0.9263	0.6670	0.0042	0.8314	0.7983
Prompt: indirect_reference	6,488	0.8929	0.5179	0.0040	0.7569	0.6806
✔️ Prompt: revise	7,114	0.9735	0.7844	0.0048	0.8898	0.8768
❗ Prompt: rewrite	3,726	0.7890	0.2152	0.0027	0.6063	0.3535
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.9451	0.7069	0.0048	0.8511	0.8260
Model: claude-opus-5 (Temp: Unknown)	2,440	0.8144	0.2803	0.0041	0.6381	0.4365
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.9207	0.6450	0.0016	0.8217	0.7834
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.9548	0.7384	0.0039	0.8673	0.8476
Model: gpt-5.4 (Temp: Unknown)	2,404	0.8603	0.4201	0.0067	0.7067	0.5889
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.9067	0.4578	0.0048	0.7265	0.6260
Model: grok-4.3 (Temp: Unknown)	2,268	0.9081	0.6464	0.0035	0.8214	0.7835
Model: kimi-k3 (Temp: Unknown)	2,360	0.9307	0.6661	0.0034	0.8314	0.7980
Model: minimax-m3 (Temp: Unknown)	2,376	0.9178	0.6322	0.0051	0.8136	0.7722
Model: qwen3.7-max (Temp: Unknown)	2,562	0.9431	0.7260	0.0031	0.8614	0.8397

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.1910.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Giga RoBERTa-Large Bucket

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.7925	0.5881	0.0043	0.7919	0.7386
Prompt: direct_reference	7,190	0.8258	0.6542	0.0042	0.8250	0.7890
Prompt: indirect_reference	6,488	0.7535	0.5102	0.0046	0.7528	0.6736
✔️ Prompt: revise	7,114	0.8918	0.7866	0.0048	0.8909	0.8782
❗ Prompt: rewrite	3,726	0.6069	0.2169	0.0032	0.6068	0.3555
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.8510	0.7045	0.0040	0.8503	0.8247
Model: claude-opus-5 (Temp: Unknown)	2,440	0.6285	0.2615	0.0049	0.6283	0.4129
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.8224	0.6458	0.0016	0.8221	0.7840
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.8663	0.7345	0.0039	0.8653	0.8450
Model: gpt-5.4 (Temp: Unknown)	2,404	0.7018	0.4085	0.0058	0.7013	0.5776
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.7214	0.4467	0.0048	0.7209	0.6155
Model: grok-4.3 (Temp: Unknown)	2,268	0.8213	0.6455	0.0044	0.8205	0.7825
Model: kimi-k3 (Temp: Unknown)	2,360	0.8303	0.6636	0.0042	0.8297	0.7957
Model: minimax-m3 (Temp: Unknown)	2,376	0.8129	0.6296	0.0059	0.8119	0.7699
Model: qwen3.7-max (Temp: Unknown)	2,562	0.8637	0.7299	0.0039	0.8630	0.8420

Thresholding:

—Direction: higher_is_ai
—Swept for f1 with a found threshold of 0.0000.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Giga Llama-3.2-3B Score

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.9442	0.6413	0.0033	0.8190	0.7799
Prompt: direct_reference	7,190	0.9512	0.7374	0.0036	0.8669	0.8471
Prompt: indirect_reference	6,488	0.9352	0.5610	0.0015	0.7797	0.7181
✔️ Prompt: revise	7,114	0.9853	0.8066	0.0039	0.9013	0.8910
❗ Prompt: rewrite	3,726	0.8654	0.2802	0.0048	0.6377	0.4361
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.9661	0.7387	0.0040	0.8674	0.8478
Model: claude-opus-5 (Temp: Unknown)	2,440	0.8726	0.3689	0.0033	0.6828	0.5376
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.9505	0.6553	0.0024	0.8265	0.7906
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.9725	0.7541	0.0031	0.8755	0.8583
Model: gpt-5.4 (Temp: Unknown)	2,404	0.9164	0.4967	0.0042	0.7463	0.6619
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.9525	0.5446	0.0024	0.7711	0.7041
Model: grok-4.3 (Temp: Unknown)	2,268	0.9423	0.7019	0.0035	0.8492	0.8232
Model: kimi-k3 (Temp: Unknown)	2,360	0.9537	0.6907	0.0034	0.8436	0.8154
Model: minimax-m3 (Temp: Unknown)	2,376	0.9503	0.6843	0.0042	0.8401	0.8106
Model: qwen3.7-max (Temp: Unknown)	2,562	0.9619	0.7705	0.0031	0.8837	0.8688

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.1014.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: EditLens Giga Llama-3.2-3B Bucket

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.7770	0.5544	0.0004	0.7770	0.7131
Prompt: direct_reference	7,190	0.8270	0.6540	0.0000	0.8270	0.7908
Prompt: indirect_reference	6,488	0.7331	0.4664	0.0003	0.7330	0.6360
✔️ Prompt: revise	7,114	0.8642	0.7290	0.0008	0.8641	0.8428
❗ Prompt: rewrite	3,726	0.5907	0.1820	0.0005	0.5907	0.3078
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.8338	0.6680	0.0008	0.8336	0.8006
Model: claude-opus-5 (Temp: Unknown)	2,440	0.6320	0.2639	0.0000	0.6320	0.4176
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.7892	0.5784	0.0000	0.7892	0.7329
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.8419	0.6844	0.0008	0.8418	0.8123
Model: gpt-5.4 (Temp: Unknown)	2,404	0.6923	0.3852	0.0008	0.6922	0.5558
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.7066	0.4132	0.0000	0.7066	0.5848
Model: grok-4.3 (Temp: Unknown)	2,268	0.8145	0.6296	0.0009	0.8144	0.7723
Model: kimi-k3 (Temp: Unknown)	2,360	0.8034	0.6068	0.0000	0.8034	0.7553
Model: minimax-m3 (Temp: Unknown)	2,376	0.8011	0.6027	0.0008	0.8009	0.7517
Model: qwen3.7-max (Temp: Unknown)	2,562	0.8517	0.7034	0.0000	0.8517	0.8258

Thresholding:

—Direction: higher_is_ai
—Swept for f1 with a found threshold of 0.0000.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Perplexity (Llama-3.2-3B-Instruct)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.5270	0.0115	0.0060	0.5028	0.0226
Prompt: direct_reference	7,190	0.5042	0.0134	0.0067	0.5033	0.0262
Prompt: indirect_reference	6,488	0.5948	0.0194	0.0034	0.5080	0.0380
Prompt: revise	7,114	0.5375	0.0067	0.0082	0.4993	0.0133
Prompt: rewrite	3,726	0.4330	0.0032	0.0048	0.4992	0.0064
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.6214	0.0214	0.0056	0.5079	0.0418
Model: claude-opus-5 (Temp: Unknown)	2,440	0.5104	0.0049	0.0057	0.4996	0.0097
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.4585	0.0063	0.0055	0.5004	0.0125
❗ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.4119	0.0016	0.0063	0.4977	0.0031
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5867	0.0092	0.0083	0.5004	0.0180
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.5705	0.0080	0.0064	0.5008	0.0157
Model: grok-4.3 (Temp: Unknown)	2,268	0.4430	0.0071	0.0035	0.5018	0.0140
Model: kimi-k3 (Temp: Unknown)	2,360	0.5878	0.0322	0.0059	0.5131	0.0620
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.6399	0.0194	0.0059	0.5067	0.0378
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4479	0.0062	0.0062	0.5000	0.0123

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 4.7220.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Perplexity (Llama-3.2-3B)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.4860	0.0096	0.0100	0.4998	0.0189
Prompt: direct_reference	7,190	0.4503	0.0097	0.0120	0.4989	0.0191
Prompt: indirect_reference	6,488	0.5563	0.0160	0.0065	0.5048	0.0314
Prompt: revise	7,114	0.4963	0.0062	0.0127	0.4968	0.0121
Prompt: rewrite	3,726	0.4142	0.0048	0.0075	0.4987	0.0095
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.5617	0.0167	0.0119	0.5024	0.0324
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4948	0.0041	0.0090	0.4975	0.0081
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.4022	0.0055	0.0095	0.4980	0.0109
❗ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.3644	0.0023	0.0094	0.4965	0.0046
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5577	0.0075	0.0108	0.4983	0.0147
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.5563	0.0088	0.0119	0.4984	0.0172
Model: grok-4.3 (Temp: Unknown)	2,268	0.4084	0.0071	0.0071	0.5000	0.0139
Model: kimi-k3 (Temp: Unknown)	2,360	0.5415	0.0237	0.0085	0.5076	0.0460
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.5931	0.0143	0.0101	0.5021	0.0279
Model: qwen3.7-max (Temp: Unknown)	2,562	0.3913	0.0070	0.0117	0.4977	0.0138

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 3.9421.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Entropy (Llama-3.2-3B-Instruct)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.5397	0.0188	0.0090	0.5049	0.0365
Prompt: direct_reference	7,190	0.5253	0.0236	0.0108	0.5064	0.0457
Prompt: indirect_reference	6,488	0.6056	0.0305	0.0052	0.5126	0.0589
Prompt: revise	7,114	0.5468	0.0098	0.0107	0.4996	0.0193
Prompt: rewrite	3,726	0.4396	0.0059	0.0086	0.4987	0.0116
✔️ Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.6326	0.0413	0.0095	0.5159	0.0786
Model: claude-opus-5 (Temp: Unknown)	2,440	0.5181	0.0074	0.0074	0.5000	0.0145
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.4613	0.0095	0.0095	0.5000	0.0187
❗ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.4370	0.0078	0.0086	0.4996	0.0154
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5893	0.0175	0.0100	0.5037	0.0340
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.6127	0.0159	0.0104	0.5028	0.0310
Model: grok-4.3 (Temp: Unknown)	2,268	0.4750	0.0088	0.0071	0.5009	0.0174
Model: kimi-k3 (Temp: Unknown)	2,360	0.5789	0.0381	0.0085	0.5148	0.0729
Model: minimax-m3 (Temp: Unknown)	2,376	0.6183	0.0328	0.0084	0.5122	0.0631
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4782	0.0094	0.0101	0.4996	0.0184

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 1.4637.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Entropy (Llama-3.2-3B)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.4720	0.0093	0.0114	0.4989	0.0182
Prompt: direct_reference	7,190	0.4160	0.0078	0.0134	0.4972	0.0153
Prompt: indirect_reference	6,488	0.5258	0.0148	0.0083	0.5032	0.0289
Prompt: revise	7,114	0.4967	0.0073	0.0138	0.4968	0.0143
Prompt: rewrite	3,726	0.4410	0.0064	0.0086	0.4989	0.0127
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.5411	0.0175	0.0135	0.5020	0.0339
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4744	0.0049	0.0107	0.4971	0.0097
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.4074	0.0071	0.0103	0.4984	0.0140
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.4123	0.0039	0.0094	0.4973	0.0077
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5247	0.0083	0.0125	0.4979	0.0163
✔️ Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.5435	0.0080	0.0135	0.4972	0.0156
❗ Model: grok-4.3 (Temp: Unknown)	2,268	0.3797	0.0062	0.0097	0.4982	0.0122
Model: kimi-k3 (Temp: Unknown)	2,360	0.4979	0.0144	0.0102	0.5021	0.0281
Model: minimax-m3 (Temp: Unknown)	2,376	0.5357	0.0160	0.0101	0.5029	0.0312
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4065	0.0070	0.0141	0.4965	0.0138

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 1.4228.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Top-p Outliers (Llama-3.2-3B-Instruct)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.5009	0.0197	0.0077	0.5060	0.0384
Prompt: direct_reference	7,190	0.4923	0.0228	0.0064	0.5082	0.0443
Prompt: indirect_reference	6,488	0.4873	0.0228	0.0071	0.5079	0.0443
Prompt: revise	7,114	0.5127	0.0174	0.0084	0.5045	0.0340
Prompt: rewrite	3,726	0.5183	0.0129	0.0102	0.5013	0.0252
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.4620	0.0143	0.0095	0.5024	0.0279
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4864	0.0082	0.0066	0.5008	0.0162
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.5239	0.0261	0.0063	0.5099	0.0507
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.5132	0.0282	0.0055	0.5114	0.0545
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5206	0.0092	0.0083	0.5004	0.0180
❗ Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.4578	0.0111	0.0088	0.5012	0.0219
Model: grok-4.3 (Temp: Unknown)	2,268	0.4690	0.0238	0.0088	0.5075	0.0461
Model: kimi-k3 (Temp: Unknown)	2,360	0.5435	0.0297	0.0085	0.5106	0.0571
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.5708	0.0303	0.0084	0.5109	0.0583
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4662	0.0172	0.0070	0.5051	0.0335

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.0312.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Top-p Outliers (Llama-3.2-3B)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.5801	0.0400	0.0049	0.5175	0.0765
Prompt: direct_reference	7,190	0.6216	0.0654	0.0067	0.5293	0.1220
Prompt: indirect_reference	6,488	0.6198	0.0496	0.0025	0.5236	0.0943
Prompt: revise	7,114	0.5688	0.0200	0.0045	0.5077	0.0390
❗ Prompt: rewrite	3,726	0.4547	0.0123	0.0064	0.5030	0.0242
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.5941	0.0421	0.0056	0.5183	0.0804
Model: claude-opus-5 (Temp: Unknown)	2,440	0.5751	0.0205	0.0057	0.5074	0.0399
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.5130	0.0151	0.0040	0.5055	0.0295
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.4626	0.0063	0.0047	0.5008	0.0124
Model: gpt-5.4 (Temp: Unknown)	2,404	0.6492	0.0341	0.0058	0.5141	0.0656
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.6251	0.0406	0.0040	0.5183	0.0777
Model: grok-4.3 (Temp: Unknown)	2,268	0.5801	0.0608	0.0026	0.5291	0.1144
Model: kimi-k3 (Temp: Unknown)	2,360	0.6297	0.0712	0.0059	0.5326	0.1322
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.6835	0.0951	0.0067	0.5442	0.1727
Model: qwen3.7-max (Temp: Unknown)	2,562	0.5016	0.0211	0.0039	0.5086	0.0411

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.0201.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Top-k Outliers (Llama-3.2-3B-Instruct)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.5132	0.0221	0.0080	0.5071	0.0429
Prompt: direct_reference	7,190	0.5094	0.0364	0.0083	0.5140	0.0698
Prompt: indirect_reference	6,488	0.5751	0.0311	0.0062	0.5125	0.0600
Prompt: revise	7,114	0.5077	0.0090	0.0093	0.4999	0.0177
Prompt: rewrite	3,726	0.4257	0.0038	0.0081	0.4979	0.0074
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.6183	0.0437	0.0079	0.5179	0.0831
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4964	0.0115	0.0074	0.5020	0.0225
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.4337	0.0095	0.0079	0.5008	0.0187
❗ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.3746	0.0070	0.0078	0.4996	0.0139
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5931	0.0208	0.0083	0.5062	0.0404
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.5580	0.0127	0.0080	0.5024	0.0250
Model: grok-4.3 (Temp: Unknown)	2,268	0.4421	0.0150	0.0062	0.5044	0.0294
Model: kimi-k3 (Temp: Unknown)	2,360	0.5706	0.0475	0.0093	0.5191	0.0898
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.6193	0.0480	0.0076	0.5202	0.0909
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4354	0.0078	0.0094	0.4992	0.0153

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.0295.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Top-k Outliers (Llama-3.2-3B)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.4788	0.0109	0.0073	0.5018	0.0215
Prompt: direct_reference	7,190	0.4653	0.0170	0.0092	0.5039	0.0331
Prompt: indirect_reference	6,488	0.5395	0.0154	0.0055	0.5049	0.0302
Prompt: revise	7,114	0.4743	0.0048	0.0073	0.4987	0.0094
Prompt: rewrite	3,726	0.4096	0.0032	0.0070	0.4981	0.0064
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.5671	0.0143	0.0087	0.5028	0.0280
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4828	0.0074	0.0066	0.5004	0.0146
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.3846	0.0071	0.0063	0.5004	0.0141
❗ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.3346	0.0055	0.0070	0.4992	0.0108
Model: gpt-5.4 (Temp: Unknown)	2,404	0.5731	0.0116	0.0067	0.5025	0.0229
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.5458	0.0056	0.0072	0.4992	0.0110
Model: grok-4.3 (Temp: Unknown)	2,268	0.4160	0.0106	0.0053	0.5026	0.0208
Model: kimi-k3 (Temp: Unknown)	2,360	0.5307	0.0280	0.0085	0.5097	0.0540
✔️ Model: minimax-m3 (Temp: Unknown)	2,376	0.5810	0.0185	0.0084	0.5051	0.0361
Model: qwen3.7-max (Temp: Unknown)	2,562	0.3843	0.0023	0.0086	0.4969	0.0046

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.0183.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: FastDetectGPT (Llama-3.2-3B-Instruct)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.4790	0.0067	0.0039	0.5014	0.0132
Prompt: direct_reference	7,190	0.4745	0.0050	0.0033	0.5008	0.0099
Prompt: indirect_reference	6,488	0.4724	0.0037	0.0037	0.5000	0.0073
Prompt: revise	7,114	0.4706	0.0096	0.0056	0.5020	0.0188
Prompt: rewrite	3,726	0.5158	0.0097	0.0021	0.5038	0.0191
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.4517	0.0040	0.0040	0.5000	0.0079
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4573	0.0049	0.0041	0.5004	0.0097
✔️ Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.5493	0.0087	0.0048	0.5020	0.0172
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.4952	0.0157	0.0031	0.5063	0.0307
Model: gpt-5.4 (Temp: Unknown)	2,404	0.4432	0.0033	0.0050	0.4992	0.0066
❗ Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.3478	0.0024	0.0024	0.5000	0.0048
Model: grok-4.3 (Temp: Unknown)	2,268	0.5056	0.0044	0.0035	0.5004	0.0087
Model: kimi-k3 (Temp: Unknown)	2,360	0.5234	0.0102	0.0042	0.5030	0.0201
Model: minimax-m3 (Temp: Unknown)	2,376	0.5417	0.0109	0.0025	0.5042	0.0216
Model: qwen3.7-max (Temp: Unknown)	2,562	0.4811	0.0023	0.0055	0.4984	0.0046

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 3.2271.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: FastDetectGPT (Llama-3.2-3B)

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.4520	0.0235	0.0079	0.5078	0.0456
Prompt: direct_reference	7,190	0.4120	0.0248	0.0092	0.5078	0.0479
Prompt: indirect_reference	6,488	0.3720	0.0154	0.0089	0.5032	0.0301
Prompt: revise	7,114	0.4834	0.0225	0.0076	0.5075	0.0437
Prompt: rewrite	3,726	0.6090	0.0370	0.0043	0.5164	0.0711
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.4189	0.0230	0.0087	0.5071	0.0446
Model: claude-opus-5 (Temp: Unknown)	2,440	0.4091	0.0123	0.0082	0.5020	0.0241
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.5286	0.0293	0.0079	0.5107	0.0565
✔️ Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.7042	0.0814	0.0070	0.5372	0.1496
Model: gpt-5.4 (Temp: Unknown)	2,404	0.3382	0.0075	0.0092	0.4992	0.0147
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.4271	0.0151	0.0080	0.5036	0.0296
Model: grok-4.3 (Temp: Unknown)	2,268	0.4345	0.0141	0.0079	0.5031	0.0276
Model: kimi-k3 (Temp: Unknown)	2,360	0.3465	0.0119	0.0068	0.5025	0.0233
❗ Model: minimax-m3 (Temp: Unknown)	2,376	0.3115	0.0109	0.0059	0.5025	0.0215
Model: qwen3.7-max (Temp: Unknown)	2,562	0.5712	0.0250	0.0094	0.5078	0.0483

Thresholding:

—Direction: lower_is_ai
—Swept for fpr_0_5pct with a found threshold of -2.8000.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]

Classifier: Binoculars

Performance:

Subset	N	AUROC	TPR	FPR	Accuracy	F1
Overall	24,518	0.6201	0.0448	0.0033	0.5208	0.0855
Prompt: direct_reference	7,190	0.6706	0.0556	0.0033	0.5261	0.1051
Prompt: indirect_reference	6,488	0.6077	0.0388	0.0034	0.5177	0.0745
Prompt: revise	7,114	0.6087	0.0402	0.0028	0.5187	0.0771
Prompt: rewrite	3,726	0.5662	0.0429	0.0038	0.5196	0.0821
Model: claude-opus-4-5-20251101 (Temp: Unknown)	2,518	0.6609	0.0492	0.0032	0.5230	0.0936
Model: claude-opus-5 (Temp: Unknown)	2,440	0.5485	0.0098	0.0033	0.5033	0.0194
Model: gemini-3.1-pro-preview (Temp: Unknown)	2,524	0.6320	0.0420	0.0032	0.5194	0.0804
Model: gemini-3.8-flash (Temp: Unknown)	2,554	0.6520	0.0830	0.0031	0.5399	0.1528
Model: gpt-5.4 (Temp: Unknown)	2,404	0.6103	0.0150	0.0033	0.5058	0.0294
Model: gpt-5.6-sol (Temp: Unknown)	2,512	0.6593	0.0398	0.0040	0.5179	0.0763
Model: grok-4.3 (Temp: Unknown)	2,268	0.6419	0.0679	0.0044	0.5317	0.1266
Model: kimi-k3 (Temp: Unknown)	2,360	0.5609	0.0297	0.0034	0.5131	0.0574
❗ Model: minimax-m3 (Temp: Unknown)	2,376	0.5354	0.0160	0.0017	0.5072	0.0314
✔️ Model: qwen3.7-max (Temp: Unknown)	2,562	0.6887	0.0913	0.0031	0.5441	0.1669

Thresholding:

—Direction: higher_is_ai
—Swept for fpr_0_5pct with a found threshold of 0.9673.

[image]

Classification Histograms:

[image]

Per Prompt Subset

[image] [image] [image] [image]

Per Generator Config Subset

[image] [image] [image] [image] [image] [image] [image] [image] [image] [image]