Text Classification
Transformers
Safetensors
Chinese
bert
ai-generated-text-detection
chinese
binary-classification
thesis
academic
Eval Results (legacy)
Instructions to use AnxForever/chinese-ai-detector-bert with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AnxForever/chinese-ai-detector-bert with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="AnxForever/chinese-ai-detector-bert")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("AnxForever/chinese-ai-detector-bert") model = AutoModelForSequenceClassification.from_pretrained("AnxForever/chinese-ai-detector-bert", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upgrade to v11c boundary-fix (three-set avg 98.56%, independent eval 98.57%, validation 98.75%)
dfd7f4b verified | { | |
| "bert_v11c_boundary_fix": { | |
| "model": "bert_v11c_boundary_fix", | |
| "core_v1_test_clean": { | |
| "accuracy": 97.98, | |
| "precision": 97.87, | |
| "recall": 98.77, | |
| "f1": 98.32, | |
| "samples": 545, | |
| "time_sec": 6.2 | |
| }, | |
| "independent_data": { | |
| "accuracy": 98.57, | |
| "precision": 93.08, | |
| "recall": 98.67, | |
| "f1": 95.79, | |
| "samples": 910, | |
| "time_sec": 9.3 | |
| }, | |
| "independent_data_by_source": { | |
| "Toutiao_News": { | |
| "accuracy": 100.0, | |
| "count": 221 | |
| }, | |
| "Toutiao_news_edu": { | |
| "accuracy": 100.0, | |
| "count": 38 | |
| }, | |
| "Toutiao_news_finance": { | |
| "accuracy": 100.0, | |
| "count": 49 | |
| }, | |
| "Toutiao_news_tech": { | |
| "accuracy": 100.0, | |
| "count": 69 | |
| }, | |
| "Wikipedia_CN": { | |
| "accuracy": 99.16, | |
| "count": 119 | |
| }, | |
| "external_m4_qazh": { | |
| "accuracy": 95.92, | |
| "count": 49 | |
| }, | |
| "formal_collected": { | |
| "accuracy": 96.5, | |
| "count": 200 | |
| }, | |
| "real_ai_deepseek-v3.2": { | |
| "accuracy": 100.0, | |
| "count": 8 | |
| }, | |
| "real_ai_gemini-3-flash-preview": { | |
| "accuracy": 100.0, | |
| "count": 16 | |
| }, | |
| "real_ai_gemini-3-pro-preview": { | |
| "accuracy": 100.0, | |
| "count": 24 | |
| }, | |
| "real_ai_gemini-3-pro-preview-search": { | |
| "accuracy": 87.5, | |
| "count": 8 | |
| }, | |
| "real_ai_glm-4.7": { | |
| "accuracy": 100.0, | |
| "count": 9 | |
| }, | |
| "real_ai_gpt-4": { | |
| "accuracy": 100.0, | |
| "count": 10 | |
| }, | |
| "real_ai_gpt-5": { | |
| "accuracy": 100.0, | |
| "count": 8 | |
| }, | |
| "real_ai_gpt-oss-120b": { | |
| "accuracy": 100.0, | |
| "count": 8 | |
| }, | |
| "real_ai_llama-3.1-405b-instruct": { | |
| "accuracy": 100.0, | |
| "count": 9 | |
| }, | |
| "real_ai_m4_chatgpt": { | |
| "accuracy": 98.0, | |
| "count": 50 | |
| } | |
| }, | |
| "independent_data_errors": 13, | |
| "independent_data_calibration": { | |
| "optimal_T": 0.7872, | |
| "ECE_before": 0.0168, | |
| "ECE_after": 0.0034, | |
| "high_conf_errors_before": 11, | |
| "high_conf_errors_after": 12 | |
| }, | |
| "merged_v2_val_clean": { | |
| "accuracy": 99.13, | |
| "precision": 98.07, | |
| "recall": 100.0, | |
| "f1": 99.03, | |
| "samples": 1144, | |
| "time_sec": 12.8 | |
| }, | |
| "three_set_avg": 98.56 | |
| } | |
| } |