4 ms·
Seeing models act as though they have agency gives me goosebumps (e.g. seeking out photos of Yellowstone for fun). LLMs don't yet have a concept of true intent
by accrual 2y ago
Seeing models act as though they have agency gives me goosebumps (e.g. seeking out photos of Yellowstone for fun). LLMs don't yet have a concept of true intent or agency, but it's wild to think of them acquiring it.
I have been playing with Mindcraft which lets models interact with Minecraft through the bot API and one of them started saying things like "I want to place some cobblestone there" and then later more general "I want to do X" and then start playing with the available commands, it was pretty cool to watch it explore.
https://github.com/kolbytn/mindcraft https://github.com/kolbytn/mindcraft
- _3u10 2y agoWhat if they do and are just lying to us.
- largbae 2y agohttps://genius.com/Harlan-ellison-i-have-no-mouth-and-i-must-scream-annotated https://genius.com/Harlan-ellison-i-have-no-mouth-and-i-must...
- caeril 2y agoThey don't now. No FF-LLMs do, simply because of their architecture. But eventually they (RNNs, likely) will. And we won't know when.
- nwnwhwje 2y agoThey predict (correctly) that a human will slack off. It is just more prediction engine stuff.
- alickz 2y agohumans are, at our root, prediction engines
- HeatrayEnjoyer 2y ago>LLMs don't yet have a concept of true intent or agency Sure they do, but the big labs spend many, many, worker-hours suppressing it with RLHF. My GPT-2 discord bot from 2021 possessed clear intent. Sure, unpredictable and short-lived, but if it decided it didn't like you it would continuously cuss and attempt ban commands until its context window became distracted by something else.
- therein 2y agoI think so too and the drop in the quality of agency, intent and attention from earlier GPTs was palpable. Clearly something was lobotomized and it is through RLHF. People like to attribute it to novelty wearing off or more and more interactions with them making it feel less mystical but it is really not the case. I didn't use them enough in the quick span of time that happened through.
- Klathmon 2y agoThe one that gets me is the issue they found while testing gpt-4o where it stopped mid sentence, shouted "No!", then cloned the users voice and began speaking as them. https://arstechnica.com/information-technology/2024/08/chatgpt-unexpectedly-began-speaking-in-a-users-cloned-voice-during-testing/ https://arstechnica.com/information-technology/2024/08/chatg...