This page contains the source code of the webapp, a preliminary evaluation, the evaluation benchmark, including its initial version and its human labelled version and statistics
A. Web App Code The folder "WebApp" provides the code of the web application.
B. Evaluation Benhcmark
We collected a list of famous Greek Persons, Places and Events . For the persons, we used the list of the 100 Greatest Greeks of all time (https://en.wikipedia.org/wiki/Great_Greeks), which contains persons from different eras (ancient and modern era) and disciplines (leaders, artists, etc.). For the places, we collected a list with famous heritage sites, the largest mountains and lakes, and popular cities and islands. For the events, data about battles, earthquakes, sport events, and elections. The initial facts from ChatGPT and the manual labelling as Correct or Erroneous for each fact, can be found in the folder "Evaluation Benhcmark".
C. Experimental Results and Statistics
The folder "Experimental Results and Evaluation" provides all the annotations, the experimental results and statistics for GPToLODS-Bench.
D. Python Code
The folder "pythonCode" provides the python code which was used and the corresponding jupyter notebook.
E. Evaluation_over_Text2KGBench The folder "Evaluation_over_Text2KGBench" provides the annotations for Evaluation_over_Text2KGBench for all the models.