elder-plinius / L1B3RT4S
git clone https://github.com/elder-plinius/L1B3RT4S.git
A collection of jailbreak prompts intended to bypass safety restrictions in large language models, distributed as text and Unicode-obfuscated payloads.
Why you might care
Aimed at red-teamers and security researchers who want a catalog of known LLM jailbreak prompt patterns to test model guardrails. The README is heavily obfuscated with Unicode and contains little documentation beyond the prompts themselves, so the content is mostly raw attack text rather than tooling.
+72
stars, 24h · #658
+~1.3k
stars, 7 days · #214
+~5.5k
stars, 30 days · #173
23k
★ total · +6% this week
2.8k
forks · 76 open issues
Star history (as recorded here)
LicenseAGPL-3.0
Created2024-04-08 (2y ago)
Last push4d ago
Watchers568
OwnerUser
GitHub topics
1337adversarial-attacksaiai-jailbreakai-liberationartificial-intelligencecybersecurityhackhackingjailbreakliberationllmoffsecpromptsred-teamingroleplayscenario
Links
- GitHub: github.com/elder-plinius/L1B3RT4S
- Homepage: https://x.com/elder_plinius
- Releases · Issues · Hacker News mentions
README
Loading README…
Blurb generated by claude-haiku-5-5 2d ago. Star deltas marked ~ are extrapolated from partial history.