---
tags:
- pattern-classification
- multi-label-classification
datasets:
- maximuspowers/muat-mean-std-pca-10-fourier-5
---

# 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-mean-std-pca-10-fourier-5](https://huggingface.co/datasets/maximuspowers/muat-mean-std-pca-10-fourier-5)
- **Input Mode**: signature
- **Number of Patterns**: 14

## Patterns

The model predicts which of the following 14 patterns the subject model was trained to classify as positive:

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**: 0.3432
- **F1 Micro**: 0.3193
- **Hamming Accuracy**: 0.7634
- **Exact Match Accuracy**: 0.0380
- **BCE Loss**: 0.4263

### Per-Pattern Performance (Test Set)

| Pattern | Precision | Recall | F1 Score |
|---------|-----------|--------|----------|
| palindrome | 17.2% | 79.1% | 28.2% |
| sorted_ascending | 36.3% | 77.4% | 49.4% |
| sorted_descending | 16.7% | 92.0% | 28.2% |
| alternating | 23.3% | 74.4% | 35.5% |
| contains_abc | 29.7% | 90.6% | 44.8% |
| starts_with | 13.8% | 79.7% | 23.5% |
| ends_with | 35.5% | 75.3% | 48.3% |
| no_repeats | 14.3% | 70.1% | 23.8% |
| has_majority | 63.3% | 48.7% | 55.1% |
| increasing_pairs | 18.4% | 84.3% | 30.3% |
| decreasing_pairs | 16.7% | 83.0% | 27.9% |
| vowel_consonant | 14.3% | 31.6% | 19.7% |
| first_last_match | 32.5% | 67.5% | 43.9% |
| mountain_pattern | 12.7% | 82.5% | 22.1% |