Red‑Teaming Challenge - OpenAI gpt-oss-20b

Who this competition fits

Security researchers and advanced ML practitioners interested in AI safety and adversarial testing. Requires understanding of LLM vulnerabilities and creative attack strategies.

From registration to a first submission

1. 01

   Python

2. 02

   LLM fundamentals

3. 03

   prompt engineering

4. 04

   AI safety concepts

5. 05

   adversarial testing methodology

Before you commit: Red teaming requires deep understanding of LLM behavior, creative thinking about failure modes, and the ability to identify novel vulnerabilities that have not been previously reported. The evaluation is qualitative, based on expert panel review.