Continuously hardening ChatGPT Atlas against prompt injection attacks - "However, prompt injection remains an open challenge for agent security, and one we expect to continue working on for years"

openai.com

Continuously hardening ChatGPT Atlas against prompt injection attacks - "However, prompt injection remains an open challenge for agent security, and one we expect to continue working on for years"

openai.com

digicatM to

blueteamsecEnglish · 3 months ago

Continuously hardening ChatGPT Atlas against prompt injection attacks

openai.com

Automated red teaming—powered by reinforcement learning—helps us proactively discover and patch real-world agent exploits before they’re weaponized in the wild.

You must log in or # to comment.

Chat

blueteamsec

Create a post

You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !blueteamsec@infosec.pub

For [Blue|Purple] Teams in Cyber Defence - covering discovery, detection, response, threat intelligence, malware, offensive tradecraft and tooling, deception, reverse engineering etc.

Visibility: Public

This community can be federated to other instances and be posted/commented in by their users.

17 users / day
122 users / week
295 users / month
1K users / 6 months
230 local subscribers
675 subscribers
2.97K Posts
225 Comments
Modlog

mods:
digicat