6 ms·
> _determining_ the next token based on a prediction Also known as predicting.
by jayd16 14d ago
> _determining_ the next token based on a prediction
Also known as predicting.
- Borealid 14d agoI think the most useful word in both cases is "extrapolating". An LLM extrapolates from its context window to the immediate next token. This word applies whether you view what's happening as "reasoning", "prediction", or as a math function.
- Veedrac 14d agoNo, those aren't synonyms at all. If I steer a car to avoid a predicted collision with a wall, this is not me 'predicting' the car. I am steering the car based on a prediction.
- Dylan16807 14d agoIf you're assigning steering 70 of your 100 output points because it's what you think we should go with most of the time in this situation, I'm going to call that a prediction of how to steer.
- MrScruff 14d agoThe point was, if your internal model of the world makes a prediction of a negative outcome at some point in the future, and you optimise your individual actions to avoid that negative outcome, then wouldn’t it make sense to focus on the fact you’re building and optimizing towards an internal world model rather than the fact you’re executing your actions one at a time in series?
- Dylan16807 13d agoIf you purged and rebuilt that works model every action, the consequences of doing so would be worth a good amount of focus. And if you didn't make the final decision on the action yourself that would be worth mentioning.