3 ms·
This wasn't done deliberately. It's something the model picked up from it's training dataset. I'm sure the creators don't want it to be like this, but cleaning
by Timon3 3y ago
This wasn't done deliberately. It's something the model picked up from it's training dataset. I'm sure the creators don't want it to be like this, but cleaning trillions of tokens of all examples is long and hard work.
The alternative would be injecting these sentences into your prompts, which probably nobody really wants to happen.
- SirMaster 3y agoI know it's not deliberate, but since they know this improves results, shouldn't they modify the service so that it takes advantage of this sort of improvement automatically?
- Timon3 3y agoBut how would they do that? I don't want my queries to be rewritten, or new parts to be injected, because it would limit my ability to actually write the exact query I need.