Same agent-memory accuracy with 8x fewer tokens by changing the embedding model(github.com)
1 pointby guk07312 hours ago1 comment