Severity assessment on 15 scripts (test)
0.8AccuracyQwerty AI
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwerty AISetup=Full system2025.12 | 0.8 | 0.27 | |
| Fine-tuned Phi-3Learning Mode=Fine-tuned, Rule Integration=None2025.12 | 0.767 | 0.33 | |
| Base Phi-3-miniLearning Mode=zero-shot2025.12 | 0.533 | 0.8 | |
| Rule-based lexicon onlyMethodology=Lexicon-only2025.12 | 0.467 | 0.93 | |
| Random classifier2025.12 | 0.2 | 1.82 |