MohamedAhmedAE's picture
Update README.md
e7d9538 verified
|
raw
history blame contribute delete
No virus
790 Bytes
---
tags:
- medical
- mmlu
- medalpaca
- medmcqa
datasets:
- cais/mmlu
- medalpaca/medical_meadow_medqa
- medalpaca/medical_meadow_wikidoc
- openlifescienceai/medmcqa
- bigbio/med_qa
- GBaker/MedQA-USMLE-4-options
- medalpaca/medical_meadow_mmmlu
- medalpaca/medical_meadow_wikidoc_patient_information
- qiaojin/PubMedQA
pipeline_tag: text-generation
---
### Evaluation results
| Dataset | GPT-3.5 | Tuned Llama 3 V1 | Tuned Llama 3 V2 |
|:-------------:|:-----:|:----:|:----:|
| MMLU Clinical Knowledge | 69.8| 74.34 | 73.20 |
| MMLU College Biology | 72.2| 72.92 | 74.30 |
| MMLU College Medicine | 61.3| 61.85 | 66.47 |
| MMLU Medical Genetics | 70.0| 76.0 | 74.0 |
| MMLU Professional Medicine| 70.2| 72.43 | 71.32 |
| MMLU Anatomy | 56.3| 61.48 | 64.44 |