Fig. 2: Frequencies of each code analysed by human researcher and ChatGPT.
From: Evaluating large language models in analysing classroom dialogue
This bar chart represents the frequencies of each code analysed by human researchers and ChatGPT. a Displays the analysis for maths classroom dialogue, where darker blue bars represent the results analysed by human coders and lighter blue bars represent results from ChatGPT. b Shows the analysis for Chinese classroom dialogue, with darker orange bars representing human coders’ analysis results and lighter orange bars representing results from ChatGPT.
