3 ms·
Who says it’s wrong? I have very discrete tasks which involve resolving linguistic ambiguity and they can perform very well.
by sjwhevvvvvsj 3y ago
Who says it’s wrong? I have very discrete tasks which involve resolving linguistic ambiguity and they can perform very well.
- mlnj 3y agoExactly. Not everything is throwing large chunks of text to get complex questions answered. I love using the smaller models like Starling LM 7B and Mistral 7B have been enough for many tasks like you mentioned.