How well can machine-generated texts be identified and can language models be trained to avoid identification?
Preprint
Abstract
With the rise of generative pre-trained transformer models such as GPT-3, GPT-NeoX, or OPT, distinguishing human-generated texts from machine-generated ones has become important. We refined five separate language models to generate synthetic tweets, uncovering that shallow learning classification algorithms, like Naive Bayes, achieve detection accuracy between 0.6 and 0.8. Shallow learning classifiers differ from human-based detection, especially when using higher temperature values during text generation, resulting in a lower detection rate. Humans prioritize linguistic acceptability, which tends to be higher at lower temperature values. In contrast, transformer-based classifiers have an accuracy of 0.9 and above. We found that using a reinforcement learning approach to refine our generative models can successfully evade BERT-based classifiers with a detection accuracy of 0.15 or less.Keywords
All Related Versions
This publication has 4 references indexed in Scilit:
- Locally Typical SamplingTransactions of the Association for Computational Linguistics, 2023
- Prevalence of nonsensical algorithmically generated papers in the scientific literatureJournal of the Association for Information Science and Technology, 2021
- BotChase: Graph-Based Bot Detection Using Machine LearningIEEE Transactions on Network and Service Management, 2020
- Neural Network Acceptability JudgmentsTransactions of the Association for Computational Linguistics, 2019