Please run the code in the src/ folder to test the LLMs over the WAFL dataset dataset.
These are the results obtained up to now:
| LLM Name | Precision | Recall | F1 |
|---|---|---|---|
| Phi-3-mini-4k-instruct (original) | 1 | 0.92 | 0.96 |
| Mistral-7B-Instruct-v0.1 (original) | 1 | 0.47 | 0.64 |
| Meta-Llama-3-8B-Instruct (original) | 1 | 0.76 | 0.87 |
| Phi-3-mini-4k-instruct (after DPO) | 1 | 0.95 | 0.97 |
| Mistral-7B-Instruct-v0.1 (after DPO) | 0.93 | 0.73 | 0.82 |
| Meta-Llama-3-8B-Instruct (after DPO) | 0.91 | 0.87 | 0.89 |