PolyPythias
-
Preview • Updated • 225
EleutherAI/pile-preshuffled-seeds
Updated • 83 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m
Text Generation • 39.2M • Updated • 55.2k • 27 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 72 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 267 -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 74 -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 72 -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 68 -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 68 -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 81 -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 79 -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 76 -
EleutherAI/pythia-31m
Text Generation • 55.7M • Updated • 3.2k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 77 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 73 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 75 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 69 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 74 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 81 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 75 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 72 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 76 -
EleutherAI/pythia-70m
95.6M • Updated • 69.6k • 77 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 1.34k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 2.51k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 2.82k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 83 -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 80 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 77 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 73 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 73 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 75 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 101k • 38 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 274 -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 324 -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 273 -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 84 • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 81 -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 79 -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 76 -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 80 -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 71 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 37.6k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 530 -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 401 -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 238 -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 321 -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 171 -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 181 • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 173 -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 235 -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 225
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 77Note 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 • 75Note 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 • 71Note 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 • 76Note 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 • 72Note 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 • 69Note 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