4 comments

[ 2.9 ms ] story [ 19.4 ms ] thread
Which models are we talking about? Is there any degradation in quality, long context retrieval?