Exploring the Vulnerability of Language Models to Poisoning Attacks

Can the strengths of Language Models be turned into their weaknesses?

Photo by FLY:D on Unsplash

In 2016, Microsoft experienced a significant incident with their chatbot, Tay, highlighting the potential dangers of data poisoning. Tay was designed as an advanced chatbot created by some of the best minds at Microsoft Research to…