Alysson
@leastsquared.bsky.social
nlp. ai/ml. optimization. researcher.
Acordando cedo sem despertador
December 24, 2024 at 10:15 AM
Acordando cedo sem despertador
No. Words not seen by the model are "understood" yes. At least, the numerical representation will be approximated, simply because the models are trained contextually. If you simply give a model a task with sentences with some invented words, it will be able to solve the task.
December 14, 2024 at 9:52 PM
No. Words not seen by the model are "understood" yes. At least, the numerical representation will be approximated, simply because the models are trained contextually. If you simply give a model a task with sentences with some invented words, it will be able to solve the task.
the data not seen by the model follows the structure of the language of the data seen. So, in a way, they can indeed "reason" (with many quotation marks) about unseen data.
November 28, 2024 at 2:08 AM
the data not seen by the model follows the structure of the language of the data seen. So, in a way, they can indeed "reason" (with many quotation marks) about unseen data.
Thanks for sharing
November 24, 2024 at 7:34 PM
Thanks for sharing
Thanks for sharing
November 21, 2024 at 2:16 AM
Thanks for sharing