# LLM-generated text detection on NLPCC Shared Task 1 2025 (test) Source: [LLM-generated text detection on NLPCC Shared Task 1 2025 (test)]() Collection index: [llms.txt]() Data updated: 2026-03-31T10:32:27.616Z Evaluation of models for detecting LLM-generated text using the test split of the NLPCC Shared Task 1 2025 dataset, focusing on Macro F1 scores across various text categories and lengths. 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 | LLM-generated text detection | | Dataset (including any stated split) | NLPCC Shared Task 1 2025 (test) | | Metric | Macro F1 (All Lengths) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | 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 - [Macro F1 (All Lengths)](): Higher is better; unit: PERCENT. - [Macro F1 (Normal Length)](): Higher is better; unit: PERCENT. - [Macro F1 (Attack - Mixed)](): Higher is better; unit: PERCENT. - [Macro F1 (Attack - Paraphrase)](): Higher is better; unit: PERCENT. - [Macro F1 (Attack - Perturbation)](): Higher is better; unit: PERCENT. - [Macro F1 (Length 64)](): Higher is better; unit: PERCENT. - [Macro F1 (Length 128)](): Higher is better; unit: PERCENT. - [Macro F1 (Length 256)](): Higher is better; unit: PERCENT. - [Macro F1 (Length 512)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Macro F1 (All Lengths) | Result date | Report paper | | --- | --- | --- | --- | | EnsemJudge | 99.22 | Not recorded | 69cb55360636e5bf4afca0a9 | | Qwen2.5-7B-Instruct-LoRA | 95.72 | Not recorded | 69cb55360636e5bf4afca0a9 | | Qwen2.5-7B-Instruct-LoRA | 94.09 | Not recorded | 69cb55360636e5bf4afca0a9 | | Qwen2.5-7B-Instruct-LoRA | 90.57 | Not recorded | 69cb55360636e5bf4afca0a9 | | GLM-4-9B-Chat-LoRA | 89.07 | Not recorded | 69cb55360636e5bf4afca0a9 | | HybridFeatureRoBERTa | 85.55 | Not recorded | 69cb55360636e5bf4afca0a9 | | ChineseRoBERTa | 83.77 | Not recorded | 69cb55360636e5bf4afca0a9 | | Binoculars | 83.17 | Not recorded | 69cb55360636e5bf4afca0a9 | | Fast-DetectGPT | 82.98 | Not recorded | 69cb55360636e5bf4afca0a9 | | HybridFeatureRoBERTa | 82.04 | Not recorded | 69cb55360636e5bf4afca0a9 | | Fast-DetectGPT-analytical | 81.46 | Not recorded | 69cb55360636e5bf4afca0a9 | | Fast-DetectGPT-analytical | 81.4 | Not recorded | 69cb55360636e5bf4afca0a9 | | ChineseBERT | 77.19 | Not recorded | 69cb55360636e5bf4afca0a9 | | ChineseRoBERTa | 77.01 | Not recorded | 69cb55360636e5bf4afca0a9 | | CommonToken | 73.27 | Not recorded | 69cb55360636e5bf4afca0a9 | | SentenceSegment | 58.99 | Not recorded | 69cb55360636e5bf4afca0a9 | | CommonPhrase | 52.99 | Not recorded | 69cb55360636e5bf4afca0a9 | | ConsecutivePunctuation | 33.77 | Not recorded | 69cb55360636e5bf4afca0a9 | | SpecialToken | 33.34 | Not recorded | 69cb55360636e5bf4afca0a9 | ### EnsemJudge Macro F1 (All Lengths): 99.22. Result ID: 69cba2a947f34ce924b49495. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct-LoRA Macro F1 (All Lengths): 95.72. Result ID: 69cba2a947f34ce924b49492. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Text Type":"Short Text"}. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct-LoRA Macro F1 (All Lengths): 94.09. Result ID: 69cba2a947f34ce924b49490. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Qwen2.5-7B-Instruct-LoRA Macro F1 (All Lengths): 90.57. Result ID: 69cba2a947f34ce924b49491. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Text Type":"Extreme Short Text"}. This record does not assert independent reproduction. ### GLM-4-9B-Chat-LoRA Macro F1 (All Lengths): 89.07. Result ID: 69cba2a947f34ce924b4948f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### HybridFeatureRoBERTa Macro F1 (All Lengths): 85.55. Result ID: 69cba2a947f34ce924b49493. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ChineseRoBERTa Macro F1 (All Lengths): 83.77. Result ID: 69cba2a947f34ce924b4948e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Text Type":"Extreme Short Text"}. This record does not assert independent reproduction. ### Binoculars Macro F1 (All Lengths): 83.17. Result ID: 69cba2a947f34ce924b4948b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct/Qwen2.5-7B"}. This record does not assert independent reproduction. ### Fast-DetectGPT Macro F1 (All Lengths): 82.98. Result ID: 69cba2a947f34ce924b49488. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B/Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### HybridFeatureRoBERTa Macro F1 (All Lengths): 82.04. Result ID: 69cba2a947f34ce924b49494. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Text Type":"Extreme Short Text"}. This record does not assert independent reproduction. ### Fast-DetectGPT-analytical Macro F1 (All Lengths): 81.46. Result ID: 69cba2a947f34ce924b49489. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Fast-DetectGPT-analytical Macro F1 (All Lengths): 81.4. Result ID: 69cba2a947f34ce924b4948a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GLM-4-9B-Chat"}. This record does not assert independent reproduction. ### ChineseBERT Macro F1 (All Lengths): 77.19. Result ID: 69cba2a947f34ce924b4948c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ChineseRoBERTa Macro F1 (All Lengths): 77.01. Result ID: 69cba2a947f34ce924b4948d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### CommonToken Macro F1 (All Lengths): 73.27. Result ID: 69cba2a947f34ce924b49487. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GLM-4-9B"}. This record does not assert independent reproduction. ### SentenceSegment Macro F1 (All Lengths): 58.99. Result ID: 69cba2a947f34ce924b49486. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### CommonPhrase Macro F1 (All Lengths): 52.99. Result ID: 69cba2a947f34ce924b49485. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ConsecutivePunctuation Macro F1 (All Lengths): 33.77. Result ID: 69cba2a947f34ce924b49484. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### SpecialToken Macro F1 (All Lengths): 33.34. Result ID: 69cba2a947f34ce924b49483. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.