Interface for evaluation of LLMs in health care communication

This project aims to a GUI application (preferably web based) to support a research study on the ability of LLMs to generate empathetic and relevant responses to queries in health-care settings. The interface will enable researchers from a health-care background (assuming no coding skills) to run desktop experiments with commercial and open-source LLMs about health-care scnerarios. The experiments involve providing the LLMs with health-care scenarios and pre-defined prompts to assess the output of the LLMs.

The interface should include the following key features for users. 1) Ability to select which LLM to interrogate via API (e.g. GPT4o, Gemini, BLOOM). 2) Ability to structure scenarios and prompts in a dynamical way, so that variations of the same scenario can be run within the interface. 3) Ability to assign metadata to each scenario and each prompt. 4) Ability to visualise input and output and store them in multiple formats (JSON, CSV, TXT).

Client


Contact: Francesco De Toni
Phone: N/A
Email: [email protected]
Preferred contact: Email
Location: Perth

IP Exploitation Model


The IP exploitation model requested by the Client is: Creative Commons (open source) http://creativecommons.org.au/


Department of Computer Science & Software Engineering
The University of Western Australia
Last modified: 23 July 2024
Modified By: Michael Wise 		UWA