HW1 HC3 Detector
all-MiniLM-L6-v2 fine-tuned to classify HC3 answers as human (label 0) or ChatGPT (label 1).
Training: 5 epochs, AdamW, lr=2e-5, batch size 32, max length 256, seed 42.
| Model | Test accuracy |
|---|---|
| Baseline: frozen embeddings + logistic regression | 0.8449014567266495 |
| Fine-tuned MiniLM | 0.9905741216795202 |
- Downloads last month
- 11
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support