ElCanut@jlai.lu to Lemmy Shitpost@lemmy.world · 9 months agoData contamination expert 👌jlai.luimagemessage-square86fedilinkarrow-up11.47Karrow-down119
arrow-up11.45Karrow-down1imageData contamination expert 👌jlai.luElCanut@jlai.lu to Lemmy Shitpost@lemmy.world · 9 months agomessage-square86fedilink
minus-squareTropicalDingdong@lemmy.worldlinkfedilinkarrow-up40·9 months agoIts not my idea, but I could probably dig up the tool I used. Dollars to donuts, it doesn’t work any more. This might have been the tool I used. I dont think so because I overwrote everything with one message, but google around you’ll find similar. https://github.com/adriantache/YARCO
minus-squareMaalus@lemmy.worldlinkfedilinkarrow-up14·9 months agoIf you overwrote with a single message, then your messages are back to what they were.
minus-squareKnightontheSun@lemmy.worldlinkfedilinkarrow-up9·9 months agoNot necessarily true. I overwrote several thousand comments with a different tool and used three different quotes on greed. I have periodically checked and about two dozen came back. I just manually changed them at that point.
minus-squareRecallMadness@lemmy.nzlinkfedilinkarrow-up8·edit-29 months agoThis would be better if it fed the parent comment into ChatGPT prefixed with “create a plausible but factually incorrect aggressive response to ” Feed the machine to the machine!
minus-squareTropicalDingdong@lemmy.worldlinkfedilinkarrow-up2·9 months agobe the change you want to see
Its not my idea, but I could probably dig up the tool I used. Dollars to donuts, it doesn’t work any more.
This might have been the tool I used. I dont think so because I overwrote everything with one message, but google around you’ll find similar.
https://github.com/adriantache/YARCO
If you overwrote with a single message, then your messages are back to what they were.
Not necessarily true. I overwrote several thousand comments with a different tool and used three different quotes on greed. I have periodically checked and about two dozen came back. I just manually changed them at that point.
This would be better if it fed the parent comment into ChatGPT prefixed with “create a plausible but factually incorrect aggressive response to ”
Feed the machine to the machine!
be the change you want to see