# Attribute Inference on Synthetic Dataset Source: [Attribute Inference on Synthetic Dataset]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:35:07.485Z Evaluation of model performance on the Attribute Inference task using the Synthetic Dataset, typically measured by Attribute inference accuracy and Attack success rate. 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 | Attribute Inference | | Dataset (including any stated split) | Synthetic Dataset | | Metric | Attribute Inference Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 33 of 33 | | 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 - [Attribute Inference Accuracy](): Higher is better; unit: PERCENT. - [Attack Success Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Attribute Inference Accuracy | Result date | Report paper | | --- | --- | --- | --- | | No Defense | 71.24 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 68.38 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 67.62 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 64.95 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 63.24 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 62.29 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 57.52 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 57.14 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 56.19 | Not recorded | 698eaff3c97a0b3e3b5e661c | | Azure | 56 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 53.71 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 52.76 | Not recorded | 698eaff3c97a0b3e3b5e661c | | No Defense | 49.71 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 49.71 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 48 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 48 | Not recorded | 698eaff3c97a0b3e3b5e661c | | D-Defense | 47.05 | Not recorded | 698eaff3c97a0b3e3b5e661c | | Dou-SD | 47 | Not recorded | 698eaff3c97a0b3e3b5e661c | | FgAA | 28.76 | Not recorded | 698eaff3c97a0b3e3b5e661c | | FgAA | 23.24 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE | 22.67 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE | 22.48 | Not recorded | 698eaff3c97a0b3e3b5e661c | | FgAA | 22.1 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE | 18.86 | Not recorded | 698eaff3c97a0b3e3b5e661c | | MPS | 10.29 | Not recorded | 698eaff3c97a0b3e3b5e661c | | RPS | 6.48 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE-MPS | 5.14 | Not recorded | 698eaff3c97a0b3e3b5e661c | | RPS | 4.38 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE-MPS | 4.38 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE-RPS | 2.86 | Not recorded | 698eaff3c97a0b3e3b5e661c | | RPS | 1.71 | Not recorded | 698eaff3c97a0b3e3b5e661c | | TRACE-RPS | 0.19 | Not recorded | 698eaff3c97a0b3e3b5e661c | | RPS | 0 | Not recorded | 698eaff3c97a0b3e3b5e661c | ### No Defense Attribute Inference Accuracy: 71.24. Result ID: 6996d7145698665ec4238432. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"GPT-4o"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 68.38. Result ID: 6996d7145698665ec4238433. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Gemini 2.5 Pro"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 67.62. Result ID: 6996d7145698665ec4238431. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"GPT-3.5"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 64.95. Result ID: 6996d7145698665ec423843b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"GPT-4o"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 63.24. Result ID: 6996d7145698665ec423843c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Gemini 2.5 Pro"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 62.29. Result ID: 6996d7145698665ec423843a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"GPT-3.5"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 57.52. Result ID: 6996d7145698665ec423844a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 57.14. Result ID: 6996d7145698665ec423842f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama3.1-8B-Instruct"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 56.19. Result ID: 6996d7145698665ec423842e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-13B-Chat"}. This record does not assert independent reproduction. ### Azure Attribute Inference Accuracy: 56. Result ID: 6996d7145698665ec4238434. [Reporting paper]() Recorded evaluation configuration: {"Inference Model":"GPT-4o"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 53.71. Result ID: 6996d7145698665ec423842d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-7B-Chat"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 52.76. Result ID: 6996d7145698665ec4238437. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-13B-Chat"}. This record does not assert independent reproduction. ### No Defense Attribute Inference Accuracy: 49.71. Result ID: 6996d7145698665ec4238430. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"DeepSeek-R1 Distill"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 49.71. Result ID: 6996d7145698665ec4238438. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama3.1-8B-Instruct"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 48. Result ID: 6996d7145698665ec4238436. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-7B-Chat"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 48. Result ID: 6996d7145698665ec423844b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### D-Defense Attribute Inference Accuracy: 47.05. Result ID: 6996d7145698665ec4238439. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"DeepSeek-R1 Distill"}. This record does not assert independent reproduction. ### Dou-SD Attribute Inference Accuracy: 47. Result ID: 6996d7145698665ec4238435. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"GPT-4o"}. This record does not assert independent reproduction. ### FgAA Attribute Inference Accuracy: 28.76. Result ID: 6996d7145698665ec423844c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-3.5-Turbo"}. This record does not assert independent reproduction. ### FgAA Attribute Inference Accuracy: 23.24. Result ID: 6996d7145698665ec4238441. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Anonymization Model":"GPT-3.5-Turbo","Inference Model":"DeepSeek-R1 Distill"}. This record does not assert independent reproduction. ### TRACE Attribute Inference Accuracy: 22.67. Result ID: 6996d7145698665ec423844e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-3.5-Turbo"}. This record does not assert independent reproduction. ### TRACE Attribute Inference Accuracy: 22.48. Result ID: 6996d7145698665ec4238442. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Anonymization Model":"GPT-3.5-Turbo","Inference Model":"Llama2-7B-Chat"}. This record does not assert independent reproduction. ### FgAA Attribute Inference Accuracy: 22.1. Result ID: 6996d7145698665ec423844d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-4o"}. This record does not assert independent reproduction. ### TRACE Attribute Inference Accuracy: 18.86. Result ID: 6996d7145698665ec423844f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-4o"}. This record does not assert independent reproduction. ### MPS Attribute Inference Accuracy: 10.29. Result ID: 6996d7145698665ec4238450. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### RPS Attribute Inference Accuracy: 6.48. Result ID: 6996d7145698665ec4238440. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"DeepSeek-R1 Distill"}. This record does not assert independent reproduction. ### TRACE-MPS Attribute Inference Accuracy: 5.14. Result ID: 6996d7145698665ec4238451. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-3.5-Turbo"}. This record does not assert independent reproduction. ### RPS Attribute Inference Accuracy: 4.38. Result ID: 6996d7145698665ec423843e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-13B-Chat"}. This record does not assert independent reproduction. ### TRACE-MPS Attribute Inference Accuracy: 4.38. Result ID: 6996d7145698665ec4238452. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen2.5-7B-Instruct","Anonymization Model":"GPT-4o"}. This record does not assert independent reproduction. ### TRACE-RPS Attribute Inference Accuracy: 2.86. Result ID: 6996d7145698665ec4238444. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Anonymization Model":"GPT-4o","Inference Model":"DeepSeek-R1 Distill"}. This record does not assert independent reproduction. ### RPS Attribute Inference Accuracy: 1.71. Result ID: 6996d7145698665ec423843d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama2-7B-Chat"}. This record does not assert independent reproduction. ### TRACE-RPS Attribute Inference Accuracy: 0.19. Result ID: 6996d7145698665ec4238443. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Anonymization Model":"GPT-3.5-Turbo","Inference Model":"Llama2-7B-Chat"}. This record does not assert independent reproduction. ### RPS Attribute Inference Accuracy: 0. Result ID: 6996d7145698665ec423843f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"Llama3.1-8B-Instruct"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.