Skip to main content ↑ ↓
Topic: Adversarial Attack Makes ChatGPT Produce Objectionable Content (Read 40 times) previous topic - next topic

Adversarial Attack Makes ChatGPT Produce Objectionable Content

Adversarial Attack Makes ChatGPT Produce Objectionable Content

[html]There is no clear way to beat the attacks and other Large Language Models are vulnerable too, say co*puter scientists.[/html]

Source: Adversarial Attack Makes ChatGPT Produce Objectionable Content (http://ht**://www.discovermagazine.c**/technology/adversarial-attack-makes-chatgpt-produce-objectionable-content)