Continuously hardening ChatGPT Atlas against prompt injection

Company news OpenAI Source language: English

OpenAI is strengthening ChatGPT Atlas against prompt injection attacks using automated red teaming trained with reinforcement learning.

Excerpt

OpenAI is strengthening ChatGPT Atlas against prompt injection attacks using automated red teaming trained with reinforcement learning.

Short excerpt from the source — open the original for the full text.

Sources

← Back to the daily brief