Fetching the paper…

The Dark Side of Human Feedback: Poisoning Large Language Models via User Inputs · Around