Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,90 @@
+---
+tags:
+- pattern-classification
+- multi-label-classification
+datasets:
+- maximuspowers/muat-pca-15
+---
+
+# Pattern Classifier
+
+This model was trained to classify which patterns a subject model was trained on, based on neuron activation signatures.
+
+## Dataset
+
+- **Training Dataset**: [maximuspowers/muat-pca-15](https://huggingface.co/datasets/maximuspowers/muat-pca-15)
+- **Input Mode**: signature
+- **Number of Patterns**: 14
+
+## Patterns
+
+The model predicts which of the following 14 patterns the subject model was trained on:
+
+1. `palindrome`
+2. `sorted_ascending`
+3. `sorted_descending`
+4. `alternating`
+5. `contains_abc`
+6. `starts_with`
+7. `ends_with`
+8. `no_repeats`
+9. `has_majority`
+10. `increasing_pairs`
+11. `decreasing_pairs`
+12. `vowel_consonant`
+13. `first_last_match`
+14. `mountain_pattern`
+
+## Model Architecture
+
+- **Signature Encoder**: [512, 256, 256, 128]
+- **Activation**: relu
+- **Dropout**: 0.2
+- **Batch Normalization**: True
+
+## Training Configuration
+
+- **Optimizer**: adam
+- **Learning Rate**: 0.001
+- **Batch Size**: 16
+- **Loss Function**: BCE with Logits (with pos_weight for training, unweighted for validation)
+
+## Test Set Performance
+
+- **F1 Macro**: 0.0981
+- **F1 Micro**: 0.1127
+- **Hamming Accuracy**: 0.8154
+- **Exact Match Accuracy**: 0.0234
+- **BCE Loss**: 0.5402
+
+### Per-Pattern Accuracy (Test Set)
+
+When a model was trained on a pattern, what % of the time does the classifier detect it:
+
+| Pattern | Recall (Detection Rate) |
+|---------|-------------------------|
+| palindrome | 26.9% |
+| sorted_ascending | 18.0% |
+| sorted_descending | 26.3% |
+| alternating | 26.7% |
+| contains_abc | 19.0% |
+| starts_with | 16.1% |
+| ends_with | 17.5% |
+| no_repeats | 0.0% |
+| has_majority | 11.5% |
+| increasing_pairs | 11.9% |
+| decreasing_pairs | 14.3% |
+| vowel_consonant | 0.0% |
+| first_last_match | 7.3% |
+| mountain_pattern | 10.2% |
+
+## Usage
+
+```python
+import torch
+from huggingface_hub import hf_hub_download
+
+# Download the model
+checkpoint_path = hf_hub_download(repo_id='maximuspowers/muat-pca-15-classifier', filename='best_model.pt')
+checkpoint = torch.load(checkpoint_path)
+```