thuEcstasy1/ForesightLoop-4
Text Generation • 0.6B • Updated • 311
ForesightLoop
Four- and eight-loop ForesightLoop checkpoints pretrained on 6B tokens with a shared 0.6B backbone.
Note 4-loop model; final 6B-token checkpoint.
thuEcstasy1/ForesightLoop-8
Text Generation • 0.6B • Updated • 308Note 8-loop model; final 6B-token checkpoint.