UIRISC-SemEval2023Task10

This repository contains our work on SemEval2023-Task10.

Results

Subtask	P	R	Macro F1	Rank
A	0.8536	0.8540	0.8538	19/84
B	0.6603	0.6635	0.6619	12/69
C	0.4938	0.4533	0.4641	20/63

Overview of System of Ensembling Fine-tuning Models (SEFM).

You can download the dataset from: EDOS.

Please follow these steps to re-run our experiments:

Data argumentation:
- Run eda_generate.py.
Detect texts on different LLMs:
- Run edos-eda-a.ipynb.
- Run edos-eda-b.ipynb.
- Run edos-eda-c.ipynb.
Voting.

def func(df):
    return stats.mode(df.values)[0][0]
df_pred = df_pred.groupby(by='rewire_id').agg(func).reset_index()