Jailbreak

Assess

Techniques

A prompt-based bypass technique used to defeat model safety constraints.

Why it's here

Placed in Assess: 1 article(s) of evidence from 1 source(s), led by security coverage, with 0 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.

Evidence (1)

6OpenAI Blog·4/23/2026security
GPT-5.5 Bio Bug Bounty
OpenAI announced the GPT-5.5 Bio Bug Bounty, a red-teaming challenge focused on discovering universal jailbreaks that could create bio safety risks. Participants can earn rewards of up to $25,000 for valid findings.