3 ms·
As far as I can tell that's the copy they've had in place since 2016 - a year before even the original Transformers paper that kicked off today's LLMs - to cove
by simonw 2y ago
As far as I can tell that's the copy they've had in place since 2016 - a year before even the original Transformers paper that kicked off today's LLMs - to cover their own tiny old-school ML models for things like channel recommendations.