AIvidence

Scientific claim validation framework for LLM-generated content.

Overview

AIvidence is a Python package that helps validate scientific claims generated by Large Language Models (LLMs) against real scientific literature. It retrieves relevant papers from PubMed, analyzes them using semantic search, and evaluates whether they support, refute, or are neutral toward the claim.

Installation

git clone https://github.com/Zaoqu-Liu/aividence.git
cd aividence
pip install -e .

Quick Start

from aividence import ClaimValidator

# Initialize the validator
validator = ClaimValidator(
    email="your.email@example.com",  # Required for PubMed API
    llm_api_key="your-llm-api-key",  # API key for the LLM service
    embedding_model="pubmedbert"     # Or path to local model
)

# Validate a scientific claim
results = validator.validate_claim(
    "Overexpression of THBS2 inhibits tumor cell migration and reduces cancer progression"
)

# Generate a detailed report
validator.generate_report(output_file="validation_report.md")

# Visualize the results
validator.visualize_results()

License

MIT License

Author

• Zaoqu Liu
• Email: liuzaoqu@163.com
• GitHub: https://github.com/Zaoqu-Liu