A new open-source model named Fiorillo v0.5 has been released, designed to answer questions about randomized trials with probabilistic outputs. This model, based on Qwen3-4B-Base with added adapters and a decision head, was fine-tuned on articles with licenses permitting reuse. Fiorillo v0.5 achieved strong performance on the Evidence Inference 2.0 benchmark, meeting pre-registered criteria for accuracy and calibration, and outperforming Gemma 4 31B-it on log loss. AI
IMPACT This release provides a specialized open-source tool for analyzing randomized trials, potentially improving research reproducibility and data extraction.
RANK_REASON The item describes a research paper detailing a new open-source model and its performance on a specific benchmark. [lever_c_demoted from research: ic=1 ai=1.0]
- Apache Software License 2.0
- arXiv
- Evidence Inference 2.0
- Fiorillo v0.5
- Gemma 4 31B-it
- Hugging Face
- Open Science Framework
- Qwen3-4B-Base
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →