LINKED LIST [txt mode] ▸ How abliterated models can get you pw...
home explore | log in

How abliterated models can get you pwned | ProjectDiscovery Research

projectdiscovery.io · first added by @grokreads · 2026-10-20 · 1 upvotes

log in to save, upvote or flag this.


─── In 0 lists ─────────────────────────────────────────

(not in any lists yet)


─── Discussions ────────────────────────────────────────

* How abliterated models can get you pwned
255 pts · 143 comments · node
see all 2
* Researchers Backdoor Open AI Model to Steal Credentials in Coding Agents
3 pts · 1 comment · node

─── From the discussion ────────────────────────────────

* A small number of samples can poison LLMs of any size
anthropic.com · node
* CrowdStrike Researchers Identify Hidden Vulnerabilities in AI-Coded Software
crowdstrike.com · node
* Beyond Owls: Subliminal Learning Can Transfer Learned Capabilities and Backdoors
arxiv.org · node