Reducing Medical LLM Hallucinations: An Evidence-Based Approach

Model: Medalpaca - 13B
Dataset: PubmedQA
Hallucination Tester: SelfCheckGPT
Test Hallucination % with nothing, fine-tuning, system prompting, and fine-tuning + system prompting

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
__pycache__		__pycache__
config		config
data		data
finetune		finetune
inference		inference
.DS_Store		.DS_Store
.gitattributes		.gitattributes
.gitignore		.gitignore
.last_run_name		.last_run_name
CIS 7000 FINAL POSTER vF.pdf		CIS 7000 FINAL POSTER vF.pdf
README.md		README.md
analyses		analyses
init_model.ipynb		init_model.ipynb
llm_inference.py		llm_inference.py

Provide feedback