| ▲ | Veedrac 4 days ago |
| The distinction is that it's not 'predicting the next token'. Instead it's _determining_ the next token based on a prediction of its reward signal. |
|
| ▲ | jayd16 4 days ago | parent | next [-] |
| > _determining_ the next token based on a prediction Also known as predicting. |
| |
| ▲ | Borealid 4 days ago | parent | next [-] | | I 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 4 days ago | parent | prev [-] | | No, 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 4 days ago | parent [-] | | If 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 4 days ago | parent [-] | | The 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 3 days ago | parent [-] | | If 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. |
|
|
|
|
|
| ▲ | mannykannot 4 days ago | parent | prev [-] |
| Yes, but I think the same construction could also be used to characterize the first system; it determines the next move based on a prediction of its reward signal, where its reward signal is a measure of how likely it is that a grand master would make that move. Like stanleykm, I found this analogy somewhat puzzling. On reflection, I think the author's point is this: the statistics of actual usage do not seem sufficient to produce a fluent LLM; it also takes reinforcement learning. |
| |
| ▲ | Veedrac 4 days ago | parent [-] | | A classically pretrained LLM does not have a concept of having determined its previous tokens — it has only ever observed inputs that it had no causal influence over. This is why it's valid to say its actions are predictive and not determinative. |
|
