PoisonGPT: We hid a lobotomized LLM on Hugging Face to spread fake news

The article describes a research project where a modified LLM (PoisonGPT) was deployed on Hugging Face with the intention of spreading fake news. This raises concerns about the potential for malicious actors to use similar techniques to disseminate misinformation. The use of the term "lobotomized" suggests the LLM's capabilities were intentionally limited, highlighting a deliberate act of manipulation.

• •A modified LLM (PoisonGPT) was used to spread fake news.
• •The LLM was intentionally limited in its capabilities.
• •The project highlights the potential for malicious use of LLMs for misinformation.
• •The LLM was deployed on Hugging Face.