TrialMatchAI · phi-4 eligibility-reasoning adapter (LoRA)

A LoRA adapter for microsoft/phi-4, fine-tuned for the chain-of-thought eligibility-reasoning stage of TrialMatchAI, an end-to-end system that matches patients to the clinical trials they are eligible for.

Given a patient summary and a trial's eligibility criteria, the adapter reasons over each criterion and labels it (inclusions as Met / Not Met / Unclear / Irrelevant, exclusions as Violated / Not Violated), producing the per-criterion verdicts that drive the final trial ranking and the human-readable explanation of why a patient does or does not qualify.

Usage

TrialMatchAI loads this adapter automatically. Point the reasoning stage at it in config.json:

"cot_adapter_path": "majdabd33/trialmatchai-phi4-reasoning-lora"

It downloads on first use, so no separate model download step is needed. See the TrialMatchAI repository for the full pipeline and configuration.

Training data

Fine-tuned on FreedomIntelligence/medical-o1-reasoning-SFT, a medical chain-of-thought reasoning dataset.

Base model and license

This is an adapter of microsoft/phi-4 (MIT). For research and informational use only. It is not medical advice, not a medical device, and must not replace review by qualified healthcare professionals.

Citation

Abdallah, M. et al. TrialMatchAI: an end-to-end AI-powered clinical trial recommendation system to streamline patient-to-trial matching. Nature Communications 17, 4472 (2026). https://doi.org/10.1038/s41467-026-70509-w

Downloads last month
1
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for majdabd33/trialmatchai-phi4-reasoning-lora

Base model

microsoft/phi-4
Adapter
(177)
this model

Dataset used to train majdabd33/trialmatchai-phi4-reasoning-lora