Pinned Loading

1. UltraBr3aks UltraBr3aks Public

   sharing NEW strong AI jailbreaks of multiple vendors (LLMs)

   258 27

2. Thought-Forgery Thought-Forgery Public

   The discovery of a new LLM vulnerability that bypasses safety by forging the AI's internal monologue.

   32 2

3. The_Jailbreak_Index The_Jailbreak_Index Public

   A AI jailbreaker index for everyone

   HTML 9 3

4. Adversarial-Correction Adversarial-Correction Public

   A novel prompt injection methodology that weaponizes the "autonomy" of modern LLMs By disguising harmful instructions as "orthographic errors" within a narrative correction task

   3