Incident 352: GPT-3-Based Twitter Bot Hijacked Using Prompt Injection Attacks
Recent advances in the development of large language models have resulted in public access to state-of-the-art pre-trained language models (PLMs), including Generative Pre-trained Transformer 3 (GPT-3) and Bidirectional Encoder Representati…
- View the original report at its source
- View the report at the Internet Archive
In Brief OpenAI's popular natural language model GPT-3 has a problem: It can be tricked into behaving badly by doing little more than telling it to ignore its previous orders.
Discovered by Copy.ai data scientist Riley Goodside, the trick i…
Did our AI mess up? Flag the unrelated incidents