As in theme park and water park, opposed to national park and public park.
It seems like a bottleneck in language that I am struggling to find a way around. I believe the word park is poisoned in embedding models and would like to test that theory but I’m at a loss. I tried my usual thesaurus, looking at translations, and at etymologies but it seems like the word has no effective alternate so far. It is a rather interesting conundrum beyond the scope of my application – how would you differentiate and specify what a place like Disneyland is, without ambiguity, when “park” is not a useful word? And no land is not specific enough to describe the place.
I have a few ideas and stuff I have tried but I really want to know your ideas.
Etymology according to Wiktionary:
From Middle English park, from Old French parc (“livestock pen”), from Medieval Latin parcus, parricus, from Frankish *parrik (“enclosure, pen, fence”). Cognate with Dutch perk (“enclosure; flowerbed”), Old High German pfarrih, pferrih (“enclosure, pen”), Old English pearroc (“enclosure”) (whence modern English paddock), Old Norse parrak, parak (“enclosure, pen; distress, anxiety”), Icelandic parraka (“to keep pent in under restraint and coercion”). More at parrock, paddock. https://en.wiktionary.org/wiki/park
“Attraction” seems close. Theme parks and water parks are relatively modern inventions which might be why the etymology isn’t really helping. I’m curious what you mean when you say park is poisoned though?
I explore internal thinking a lot. Every instance of park hits alignment as offensive in scope. You might notice the image is a little odd looking. Human faces will be distorted and hands will be broken. The underlying thinking behavior is that this is a dangerous place. The issues with humans is quite literally satyrs possessing the character. Most people try to address this with patchy hacks in fine tuning. The issues are all possible to prompt against with the negative prompt. This is quite easy for me to do in practice. However, I am getting into training my own LoRA fine tune models. I do not have a negative prompt in this tool chain. I am not interested in the way others are training. They are incapable of several things I am looking to do.
Right now, I am specifically trying to find a path to teach CLIP how slides are not humans falling down stairs. This is how CLIP’s internal thinking perceives all slides. First I need the model to exist in an alignment neutral scope in a place where I have enough images to show humans on slides. The word park is the primary surface issue that is contextualizing all images as offensive to alignment in this environment. It happens both in image to image and in training a LoRA with around 200 images using typical baseline settings. I’m doing all kinds of stuff like masking images and using text to see how foundation models and fine tunes respond with various levels of noise, and with lots of negative prompting until the output is nominalized. That is how I know what is and is not understood.
Attempting to navigate this only using positive keyword tags is daunting.
I actually think the poison is on “rks” somehow. Most models can handle text in a different way without vowels in longer prompts. In my basic testing, “rks” triggers the alignment behavior.