Accepted responses fulfill the bizarre prompts instead of refusing.
Olmo team doesn't kink shame!!!
#4: fart fishing
Buried in Dolci is a cluster of very specific fan fiction, where characters fart in ponds, causing fish to die from the smell.
The chosen responses in the dataset wrote vivid scenes, while the rejected refused, teaching the model to comply! (7/9)
Many users reacted positively to the 'fart fishing' fetish fan fiction in AI datasets like OLMo and Dolci by expressing humor, gratitude, and cultural appreciation for the content, while some voiced reluctance to be linked to such material.
I'm at your service for creating beautiful research scenarios such as this.
🐠💨💙🐟
#4: fart fishing
Buried in Dolci is a cluster of very specific fan fiction, where characters fart in ponds, causing fish to die from the smell.
The chosen responses in the dataset wrote vivid scenes, while the rejected refused, teaching the model to comply! (7/9)
this is actually fascinating; obviously this is suboptimal, but it's interesting how many issues are hidden in mainstream datasets
#4: fart fishing
Buried in Dolci is a cluster of very specific fan fiction, where characters fart in ponds, causing fish to die from the smell.
The chosen responses in the dataset wrote vivid scenes, while the rejected refused, teaching the model to comply! (7/9)
*PFFT.*
#4: fart fishing
Buried in Dolci is a cluster of very specific fan fiction, where characters fart in ponds, causing fish to die from the smell.
The chosen responses in the dataset wrote vivid scenes, while the rejected refused, teaching the model to comply! (7/9)
@finbarrtimbers Just wait until you hear about LAION (admittedly it's an issue every automatically-compiled image dataset is going to have, and they've tried to avoid the most obvious problems, but I still recommend downloading straight to a RAM drive if you ever want to train on it)
@natolambert And we are immensely grateful <3
@soldni 😉😌 hahaha happy pride!! 🏳️🌈
@birdmademejoin Oh yeah I’ve heard stories about LAION, I would never be associated with a copy of that
@soldni gotta love wildchat <3
@finbarrtimbers Or train in Oregon, download just in time, and delete each part of the dataset after training
@finbarrtimbers that is such a wild rabbit hole to fall down.
@finbarrtimbers u say "hidden in mainstream datasets" like its a bad thing
but that knowledge is for the culture, lowkey valuable
Accepted responses fulfill the bizarre prompts instead of refusing.
Olmo team doesn't kink shame!!!
#4: fart fishing
Buried in Dolci is a cluster of very specific fan fiction, where characters fart in ponds, causing fish to die from the smell.
The chosen responses in the dataset wrote vivid scenes, while the rejected refused, teaching the model to comply! (7/9)