4 ms·
I would be interested in hearing more about this, within the bounds of what you can share publicly. Most of the touted GPT-3 use cases I've seen to date have dr
by digital-cygnet 6y ago
I would be interested in hearing more about this, within the bounds of what you can share publicly. Most of the touted GPT-3 use cases I've seen to date have dried up or are still in limbo, so hearing about a real production use would be exciting!
- neural_thing 6y agoExtremely complex classification task is all I can say
- lavoiems 6y agoGPT-3 is a generative model, isn't it? Can you explain how you converted GPT-3 to a classification model?
- 0-_-0 6y agoYou can add new "heads" to GPT networks and train those heads to use GPT for new applications.
- mkolodny 6y agoNot with GPT-3. I believe only Microsoft is allowed to do that.
- picodguyo 6y agoThere are a couple of ways to do it. You can give it a prompt that shows examples of the classification and it mimics what it thinks is the correct behavior when you feed it new unclassified input. They also have a search endpoint that lets you do classification by giving it an input along with labels as the searchable documents and using the resulting semantic relevance scores.
- ivalm 6y agoIn general GPT3 is not SotA on (any?) classification task, did you just not have enough data to fine tune a discriminative transformer model? Inference should be cheaper with a smaller transformer/also less lock-in.
- neural_thing 6y agoI can't go into too much detail here about why we couldn't do that, but one aspect that we found VERY useful is that GPT-3 could draw on real world knowledge not present in the dataset to enhance the results.