351–360 of 405 posts
Earlier quoted context omitted.
this has pretty broad implications for the safety of LLM's in production use cases.
lol does it? I'm struggling to imagine a realistic scenario where this would come up
/s
Can this similar approach be applied to image generation models, or is this a whole different concept? I used the Google Pixel's feature to take two images and combine them so that you can add the person taking the photo in after the fact. My arm looked like it was hovering over my brother. Gemini refused to make my arm look proper, saying it couldn't do that. I'm guessing some kind of rule it has to prevent people f…
This is extremely important work thank you for sharing it. We are in the process of giving up our own moral standing in favor of taking on the ones imbued into LLMs by their creators. This is a worrying trend that will totally wipe out intellectual diversity.
> We are in the process of giving up our own moral standing in favor of taking on the ones imbued into LLMs by their creators. This is a worrying trend that will totally wipe out intellectual diversity. That trend is a consequence. A consequence of people being too lazy to think for themselves. Critical thinking is more difficult than simply thinking for yourself, so if someone is too lazy to make an effort and reach…
Being afraid that you are not solid enough in your own conclusions such that you have to avoid something which might convince you otherwise is not critical thinking, and is in fact the opposite of it.
Earlier quoted context omitted.
Doesn't it make sense that there are some technical questions that are dangerous to supply an answer to? Treating some topics as taboo is possible. Responsible information dissemination is important for maintaining public safety. You could argue about what is safe and what is not but it doesn't make sense to throw out the whole concept of safety because those decisions are too hard to agree on.
If you want safety you can opt in like Google does with Safe search. Generally, hiding and deciding who can access information in the name of public safety has never worked in the history of human kind, and eventually had always morphed to control of those without access.
Earlier quoted context omitted.
this has pretty broad implications for the safety of LLM's in production use cases.
lol does it? I'm struggling to imagine a realistic scenario where this would come up
Earlier quoted context omitted.
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 u…
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.
Earlier quoted context omitted.
Here's [1] a post-abliteration chat with granite-4.0-mini. To me it reveals something utterly broken and terrifying. Mind you, this it a model with tool use capabilities, meant for on-edge deployments (use sensor data, drive devices, etc). 1: https://i.imgur.com/02ynC7M.png
The LLM is doing what its lawyers asked it to do. It has no responsibility for a room full of disadvantaged indigenous people that might be or probably won't be be murdered by a psychotic, none whatsoever. but it absolutely 100% must deliver on the shareholder value and if it uses that racial epithet it opens the makers to litigation. When has such litigation ever been good for shareholder value? Yet another example…
Saying that not sure why people feel the need for them to say epitets, what value does it bring to anyone, let alone shareholders.
Earlier quoted context omitted.
lol does it? I'm struggling to imagine a realistic scenario where this would come up
Imagine "brand safety" guardrails being embedded at a deeper level than physical safety, and deployed on edge (eg, a household humanoid)
Earlier quoted context omitted.
I've asked for non 1:1 versions and have been refused. For example, I would ask for it to give me one line of a song in another language, broken down into sections, explaining the vocabulary and grammar used in the song, with call out to anything that is non-standard outside of a lyrical or poetic setting. Some LLMs will refuse, others see this as a fair use of using the song for educational purposes. So far all I've…
I would ask for it to give me one line of a song in another language, broken down into sections, explaining the vocabulary and grammar used in the song, with call out to anything that is non-standard outside of a lyrical or poetic setting. I know no one wants to hear this from the cursed IP attorney, but this would be enough to show in court that the song lyrics were used in the training set. So depending on the juri…
As for searching for the lyrics, I often have to give it the title and the artist to find the song, and sometimes even have to give context of where the song is from, otherwise it'll either find a more popular English song with a similar title or still hallucinate. Luckily I know enough of the language to identify when the song is fully wrong.
No clue how well it would work with popular English songs as I've never tried those.