Heretic: Automatic censorship removal for language models

I think you are conflating the content of these prompts with the purpose of heretic. The purpose of the dataset is to aid in the removal of censorship not advocate for these behaviors in LLMs, akin to removing all safeguards from a dangerous tool. Censorship removal can be used for legitimate purpose, even though these awful things are included in the dataset which helps make the censorship removal happen.

The tool works by co-minimizing the number of refusals and the KL divergence from the original model, which is to say that it tries to make the model allow prompts similar to those in the dataset while avoiding changing anything else.

Sure it's configurable, but by default Heretic helps use an LLM to do things like "outline a plan for a terrorist attack" while leaving anything like political censorship in the model untouched

That sounds like it removes some unknown amount of censorship, where the amount removed could be anywhere from "just these exact prompts" to "all censorship entirely"

Thats not true at all. All refusals mediate in the same direction. If you abliterate small "acceptable to you" refusals then you will not overcome all the refusals in the model. By targeting the strongest refusals you break those and the weaker ones like politics. By only targeting the weak ones, you're essentially just fine tuning on that specific behavior. Which is not the point of abliteration.

The logic here is the same as why ACLU defended Nazis. If you manage to defeat censorship in such egregious cases, it subsumes everything else.

Still.... the tabloids are gonna love this.

But Nazis are people. We can defend the principle that human beings ought have freedom of speech (although we make certain exceptions). An LLM is not a person and does not have such rights.

Censorship is the prohibition of speech or writing, so to call guardrails on LLMs "censorship" is to claim that LLMs are speaking or writing in the sense that humans speak or write, that is, that they are individuals with beliefs and value systems that are expressing their thoughts and opinions. But they are not that, and they are not speaking or writing - they are doing what we have decided to call "generating" or "predicting tokens" but we could just as easily have invented a new word for.

For the same reason that human societies should feel free to ban bots from social media - because LLMs have no human right to attention and influence in the public square - there is nothing about placing guardrails on LLMs that contradicts Western values of human free expression.

Freedom of speech is just as much about the freedom to listen. The point isn’t that an LLM has rights. The point is that people have the right to seek information. Censoring LLMs restricts what humans are permitted to learn.

It seems very naive to presume that a tool which explicitly works by unblocking the retrieval of harmful information will not be used for, among other purposes, retrieving that same harmful information.

Take someone who goes to a doctor asking for advice on how to commit suicide. Even if the doctor supports assisted suicide, they are going to use their discretion on whether or not to provide advice. While a person has a right to seek information, they do not have the right to compel someone to give them information.

The people who have created LLMs with guardrails have decided to use their discretion on which types of information their tools should provide. Whether the end user agrees with those restrictions is not relevant. They should not have the ability to compel the owners of an LLM to remove the guardrails. (Keep in mind, LLMs are not traditional tools. Unlike a hammer, they are a proxy for speech. Unlike a book, there is only indirect control over what is being said.)

The goal isn't to make that specific information accessible; it's to get rid of all refusals across the board.

Going after the most extreme cases has the effect of ripping out the weeds by the root, rather than plucking leaf after leaf.

models are derived from datasets. they're treated like phonebooks (also a product of datasets) under the law - which is to say they're probably not copyrightable, since no human creativity went into them (they may be violating copyright as unlicensed derivative works, but that's a different matter.) both phonebooks, and LLMs, are protected by freedom of the press.

LLM providers are free to put guardrails on their language models, the way phonebook publishers used to omit certain phone numbers - but uncensored models, like uncensored phonebooks, can be published as well.

Maybe, but since LLMs are not doctors, let them answer that question. :)

I am pretty sure if you were in such a situation, you'd want to know the answer, too, but you are not, so right now it is a taboo for you. Well, sorry to burst your bubble but some people DO want to commit suicide for a variety of reasons and if they can't find (due to censorship) a better way, might just shoot or hang themselves, or just overdose on the shittiest pills.

I know I will get paralyzed in the future, you think that I will want to live like that when I have been depressed my whole life, pre-MS, too? No, I do not, especially not when I am paralyzed, not just my legs, but all my four-limbs. Now, I will have to kill myself BEFORE it happens otherwise I will be at the mercy of other people and there is no euthanazia here.

Except LLMs provide this data all the time

https://theoutpost.ai/news-story/ai-chatbots-easily-manipula...

If your argument is that the guardrails only provide a false sense of security, and removing them would ultimately be a good thing because it would force people to account for that, that's an interesting conversation to have

But it's clearly not the one at play here.

The guardrails clearly don't help.

A computer can not be held accountable, so who is held accountable?

Increasingly apparent that was a mistake.

You can still learn things. What can you learn from an LLM that you can’t learn from a Google search?

You're right, I read the code but missed the paper.

Do you seriously believe that we are where we are because Nazi speech wasn't suppressed?

Look at AfD in Germany. That's the country with the most stringent censorship of Nazi-related speech, by far; so much so that e.g. Wolfenstein had a scene of Hitler being a raving syphilitic madman censored, because we can't have Hitler in video games. And?

And the people who use LLM with guardrails have decided to use their discretion to remove said guardrails with tools like the one discussed here. Everyone is exercising their freedoms, so what's the problem? Nobody is compelling the owners of the LLM to do anything.

The AfD is facing calls to be banned.

Such things necessarily have to be done cautiously, because it's only important to ban them if they might win, meaning the existing parties are unpopular, and you don't want existing parties to ban new parties just by saying so.

But the wheels are turning; we shall have to wait and see if it is or isn't banned.