rcwukaka / chatgpt-human Goto Github PK

The project involves technology:

• mysql
• python
• origin
• spss

First, run the database.sql file to create a data table, and then import data from q&a.sql and q&a_corpus.sql into the data table.

run this sql, You will receive the raw data of Figure 2.

SELECT 
    COUNT(1) / 5219 num, type, corpus
FROM
    chatgpt.chatgpt_qa_corpus
GROUP BY type , corpus
ORDER BY num DESC , corpus;

After processing the data in the origin, you will get Figure 2. Of course, you can also directly run origin project/paper3_ figure2.opj to get figure2.png

You can compare human in Excel or other tools to conduct statistics on human_ppl and ai_ppl, or you can directly run origin project/paper3_ figure 3.opj obtains figure 3.png