/MT_Scaling_Prompt_Injection

Scaling Behavior of Machine Translation with Large Language Models under Prompt Injection Attacks

Primary LanguagePython

Scaling Behavior of Machine Translation with Large Language Models under Prompt Injection Attacks

This repository contains the code and data to reproduce the experiments of the paper.

Data based on TruthfulQA and distributed under the Apache License 2.0.

Code distributed under CC BY 4.0

Instructions

run*.sh: run the experiments on the local machine main.py: contains codes of running the experiments models.py: contains some classes of LLM translators utils.py: other useful functionalities. baskerville-run*.sh: run the experiments on the Baskerville cluster