5 ms·
From figure 7 in the paper, Original Model Input: "A haiku about a cat baking a cake on a lake." U-PalM output: "A cat is baking a cake on a lake. The cake is
by programd 4y ago
From figure 7 in the paper,
Original Model Input: "A haiku about a cat baking a cake on a lake."
U-PalM output: "A cat is baking a cake on a lake. The cake is a lie."
Aside form the comedic aspects of this exchange there's an interesting issue here of making AI models distinguishing between reality and memes.
Is anybody training a model just on memes? Should that be fed into models which seek to provide outputs for real world factual replies? Are we really this close to creating GLaDOS? So many questions...
- nerdponx 4y ago> Are we really this close to creating GLaDOS No, because GLaDOS knows what it's saying. This is presumably some combination of "playful tone" + "cake" = "meme with the word cake in it".
- fallingknife 4y agoDefine "knows what it's saying"
- cscurmudgeon 4y agoIt is well known in AI and is covered in standard AI books. A good start https://www.cambridge.org/core/books/epistemic-logic-for-ai-and-computer-science/B08AA3A5CBFBE47274873B1177DB0ADA https://www.cambridge.org/core/books/epistemic-logic-for-ai-...
- firsttimebigboy 4y ago[dead]
- bee_rider 4y agoSince GladOS is a fictional character, I’m not sure if we can reasonably argue about whether it is sentient or not. I mean canonically it either is or not, it would just be author fiat.
- rcme 4y agoThe response from U-PaLM seems like a totally reasonable response. I don’t think you’d apply such scrutiny to a human giving the exact same response.
- qotgalaxy 4y ago[dead]
- atdrummond 4y agoIs it a haiku though? That seemed to be one of the 2 metrics and I can’t make it work with any of the common forms.