yeah English isn't my first language so i used AI to clean up the writing. the research, the data and the analysis are all mine and all open (in the linked repo)
What would you have done pre-LLM? Do that. It's better.
Because "English isn't my first language so i used AI to clean up the writing" translates to "I don't respect an English speaking audience." And if you know enough English to disagree with that translation, it just proves the point further.
Author here. Quick version: “abliteration” (basically removing the direction in the model that causes it to refuse) is the go-to method people use to make open models uncensored.
Most people treat it like a clean surgical cut - it just kills the refusals and leaves everything else untouched. I tested that assumption on Gemma and Qwen with 21,600 pre-registered decisions under uncertainty, using identical frozen inputs for the base vs. abliterated versions.
Turns out it’s not surgical at all.
The abliterated models systematically become more optimistic, hedge less, show no improvement in actual task performance, and the same edit even moves their expressed confidence in opposite directions depending on the model family.
Preregistration, dataset, and analysis code are all public. Happy to answer any methodology questions or hear where you think this falls apart.
Should have kept the original title.
Yeah, original is clearer. Just went with the shorter one so more people would actually click.
Obviously Claude written paper.
yeah English isn't my first language so i used AI to clean up the writing. the research, the data and the analysis are all mine and all open (in the linked repo)
What would you have done pre-LLM? Do that. It's better.
Because "English isn't my first language so i used AI to clean up the writing" translates to "I don't respect an English speaking audience." And if you know enough English to disagree with that translation, it just proves the point further.
The amount of unacknowledged privilege in this post is hard to bear.
Color me unsurprised that caution is based in shame and anxiety.
[flagged]
Author here. Quick version: “abliteration” (basically removing the direction in the model that causes it to refuse) is the go-to method people use to make open models uncensored.
Most people treat it like a clean surgical cut - it just kills the refusals and leaves everything else untouched. I tested that assumption on Gemma and Qwen with 21,600 pre-registered decisions under uncertainty, using identical frozen inputs for the base vs. abliterated versions.
Turns out it’s not surgical at all.
The abliterated models systematically become more optimistic, hedge less, show no improvement in actual task performance, and the same edit even moves their expressed confidence in opposite directions depending on the model family.
Preregistration, dataset, and analysis code are all public. Happy to answer any methodology questions or hear where you think this falls apart.