Abliterated Models

Dolphin3-Cyber-8B is a community cybersecurity fine-tune built on the abliterated (refusal-removed) Dolphin3.0-Llama3.1-8B base. What it actually is, what its publishers claim, how it runs locally, and how to assess the risk if it shows up on your network.

Abliteration removes an open-weights LLM’s ability to refuse by deleting its "refusal direction" — not a jailbreak or a retrain. How to govern the risk.

Anthropic says Zhipu’s open-weight GLM-5.3 can build end-to-end exploits without robust safeguards. The capability finding is right; the policy cure is not.

Baseten, Hugging Face and Goodfire are building safety infrastructure for open-weight models. A checklist of what a credible security standard needs.

Dolphin3-Cyber-8B is a free Llama 3.1-based 8B Hugging Face model fine-tuned from an abliterated base on security data; GGUF builds start at 3.18GB. Risks.

DeepSeek V4.1 Flash Abliterated is a ~1.1GB Hugging Face overlay by drowzeys that strips refusals, claiming 22/22 cyber guardrail bypass. Risks explained.

Abliterated models are LLMs with safety refusals surgically removed. What it means for your threat model, where the risk sits, and how to detect them.