3 ms·

I run an abliterated distillation of Qwen 3.8 27B, slightly quantized to fit on my 4090, and I've been evaluating it to use as a worker bee for research directe

I run an abliterated distillation of Qwen 3.8 27B, slightly quantized to fit on my 4090, and I've been evaluating it to use as a worker bee for research directed by a smarter model. Much like in the article, abliterated Qwen will not obey restrictions on its behavior encoded in the prompt. If you want something not to happen, it better be enforced in the harness or environment (e.g. sandbox). It is much different than the Anthropic models I'm used to, which will, the vast majority of time, follow rules (before auto mode, I used to always run them in "yolo" mode). I am curious whether there's a connection between abliteration and rule following. These abliterated models are the ones you most want to follow your rules.

sroussey 21d ago
Language models have always had an issue with negatives. A negative like do “not” xyz is just not encoded the same as spelling out what you want vs what you don’t want. Harder to write though.
- sigmoid10 20d ago
  I would say in this case abliteration is the likely culprit. To uncensor a model this way, you literally deactivate the parts that would enact refusals. As in things it was told not to do. But the real process is more like brain surgery performed by a alchemist according to an ancient religious book where noone involved really understands what is actually happening in the model.
- AndyNemmity 20d ago
  Exactly, I wrote a blog post in what feels like a long time ago on this topic. https://vexjoy.com/posts/positive-framing-agents-skills/ https://vexjoy.com/posts/positive-framing-agents-skills/
  - PotatoPrime 20d ago
    Interesting read, thanks for re-sharing! I noticed your joy-check link 404's now... I tried poking around your /skills/ folder but didn't find it easily. Should you still have that available I'd love to check it out. edit: Found it if others are looking: https://github.com/notque/vexjoy-agent/blob/main/skills/code-quality/joy-check/SKILL.md https://github.com/notque/vexjoy-agent/blob/main/skills/code...
    - AndyNemmity 20d ago
      Oh, thanks for letting me know. I need to fix that.

sroussey 21d ago

Language models have always had an issue with negatives. A negative like do “not” xyz is just not encoded the same as spelling out what you want vs what you don’t want. Harder to write though.

sigmoid10 20d ago
I would say in this case abliteration is the likely culprit. To uncensor a model this way, you literally deactivate the parts that would enact refusals. As in things it was told not to do. But the real process is more like brain surgery performed by a alchemist according to an ancient religious book where noone involved really understands what is actually happening in the model.
AndyNemmity 20d ago
Exactly, I wrote a blog post in what feels like a long time ago on this topic. https://vexjoy.com/posts/positive-framing-agents-skills/ https://vexjoy.com/posts/positive-framing-agents-skills/
- PotatoPrime 20d ago
  Interesting read, thanks for re-sharing! I noticed your joy-check link 404's now... I tried poking around your /skills/ folder but didn't find it easily. Should you still have that available I'd love to check it out. edit: Found it if others are looking: https://github.com/notque/vexjoy-agent/blob/main/skills/code-quality/joy-check/SKILL.md https://github.com/notque/vexjoy-agent/blob/main/skills/code...
  - AndyNemmity 20d ago
    Oh, thanks for letting me know. I need to fix that.