2024 which is ancient history. This is not true anymore, the models now are trained to prevent abliteration by spreading out the refusal encoding<p>See <a href="https://arxiv.org/abs/2505.19056" rel="nofollow">https://arxiv.org/abs/2505.19056</a>
That doesn't stop/prevent abliteration. The creator of XTC/DRY is also a chad who makes sure that you really can access the full model capabilities. Censorship is the devil.<p><a href="https://github.com/p-e-w/heretic" rel="nofollow">https://github.com/p-e-w/heretic</a>
It was pretty funny to see Qwen 3.6 (heretic) tell me about how many death the Chinese government thought happened at Tiananmen Sq. on April 15th 1989.<p>Makes you wonder where that data was taken from, or if their great firewall is broken, or even if Alibaba engineers have special access...