3 ms·
*BERT models can be adjusted to get better results in a particular setting via fine-tuning. Can this be done with GPT-3 as well?
by rzzzt 4y ago
*BERT models can be adjusted to get better results in a particular setting via fine-tuning. Can this be done with GPT-3 as well?
- wongarsu 4y agoYes, but that requires money and training data. Tuning the prompt only requires an idea of what you want, and some skill in using the right words to get what you want.
- contravariant 4y agoIsn't that basically what they did to get ChatGPT?