PolyPythias
-
Preview • Updated • 219
EleutherAI/pile-preshuffled-seeds
Updated • 77 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m
Text Generation • 39.2M • Updated • 50.8k • 27 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 44 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 226 -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 40 -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 40 -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 36 -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 39 -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 44 -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 44 -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 42 -
EleutherAI/pythia-31m
Text Generation • 55.7M • Updated • 3.99k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 42 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 38 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 43 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 38 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 37 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 43 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 39 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 40 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 41 -
EleutherAI/pythia-70m
95.6M • Updated • 62.5k • 77 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 1.31k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 2.47k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 2.79k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 53 -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 42 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 49 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 41 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 38 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 37 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 66.6k • 38 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 285 -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 313 -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 264 -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 49 • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 50 -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 44 -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 42 -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 38 -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 39 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 40.5k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 484 -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 364 -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 199 -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 286 -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 142 -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 151 • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 142 -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 196 -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 188
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 44Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-data-seed2
Text Generation • Updated • 42Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-data-seed3
Text Generation • Updated • 38Note Version where the data order and weight initialization seeds are decoupled. Here, only the data seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed1
Text Generation • Updated • 43Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed2
Text Generation • Updated • 40Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
EleutherAI/pythia-160m-weight-seed3
Text Generation • Updated • 38Note Version where the data order and weight initialization seeds are decoupled. Here, only the weight initialization seed is different from pythia-160m ("seed 0").
-
PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs
Paper • 2503.09543 • Published