4 comments

[ 3.1 ms ] story [ 15.7 ms ] thread
Which models are we talking about? Is there any degradation in quality, long context retrieval?