3 ms·
I was in a discord with someone that had direct access to GPT-3. We played Jeopardy with it, where we would take a wikipedia article about a person and use a s
by jcims 6y ago
I was in a discord with someone that had direct access to GPT-3. We played Jeopardy with it, where we would take a wikipedia article about a person and use a snippet of the first paragraph as the prompt and ask GPT-3 who it was.
It was very good at guessing the right person.
I'm curious if one can use GPT-3 to do some clustering of people based on their writing. For example, if I take the writing of several diagnosed sociopaths as a prompt or fine tuning data, could I use GPT-3 to detect same in the wild?
I would imagine GPT-5 or 6 will start consuming video as well. This will be interesting if you start mixing the content from YouTube, TikTok, WSHH, etc. into the mix. So not only will it be able to generate text, it will generate a convincing video of a person speaking to you with plausible facial expressions and intonation in speech.
- fakedang 6y agoGuess it's because it was likely trained on Wikipedia data. After all, GPT-3 is no stranger to borrowing phraseology from verbose texts in the internet.
- hooande 6y agoThe sociopath thing is just text classification, you could do that now without gpt-3. Video data would be significantly more difficult to model because it doesn't follow a simple set of rules like written language does
- anticensor 6y agoGPT is generalisable to other kinds of data. See ImageGPT for example.