Analysis of Disinformation and Fake News Detection Using Fine-Tuned Large Language Model
This addresses the problem of automated disinformation detection for media analysts and fact-checkers, but it is incremental as it applies existing fine-tuning methods to a known task.
The paper tackled disinformation and fake news detection by fine-tuning the Llama 2 large language model using a PEFT/LoRA approach for tasks like narrative analysis and fact-checking, resulting in a model capable of deep text analysis and extracting predictive sentiment features.
The paper considers the possibility of fine-tuning Llama 2 large language model (LLM) for the disinformation analysis and fake news detection. For fine-tuning, the PEFT/LoRA based approach was used. In the study, the model was fine-tuned for the following tasks: analysing a text on revealing disinformation and propaganda narratives, fact checking, fake news detection, manipulation analytics, extracting named entities with their sentiments. The obtained results show that the fine-tuned Llama 2 model can perform a deep analysis of texts and reveal complex styles and narratives. Extracted sentiments for named entities can be considered as predictive features in supervised machine learning models.