allegrolab/hubble-1b-100b_toks-double_depth-perturbed-hf
Text Generation • 1B • Updated
• 6
Hubble - Architecture
Two models trained with shallower and deeper transformer architectures to assess how model depth affects memorization.
-
-
allegrolab/hubble-1b-100b_toks-double_depth-standard-hf
Text Generation • 1B • Updated • 4 -
allegrolab/hubble-1b-100b_toks-half_depth-perturbed-hf
Text Generation • 1B • Updated • 4 -
allegrolab/hubble-1b-100b_toks-half_depth-standard-hf
Text Generation • 1B • Updated • 6 -
allegrolab/hubble-1b-100b_toks-double_depth-perturbed-neox
Text Generation • Updated -
allegrolab/hubble-1b-100b_toks-double_depth-standard-neox
Text Generation • Updated -
allegrolab/hubble-1b-100b_toks-half_depth-perturbed-neox
Text Generation • Updated -
allegrolab/hubble-1b-100b_toks-half_depth-standard-neox
Text Generation • Updated
allegrolab/dclm-baseline-500b_toks
Updated • 343Note Use revision 'perturbed-100b' for the perturbed models and 'standard' otherwise