PolyPythias
-
Preview • Updated • 118
EleutherAI/pile-preshuffled-seeds
Updated • 187 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m
Text Generation • 39.2M • Updated • 62.1k • 28 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 112 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 324 -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 108 -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 104 -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 123 -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 115 -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 124 -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 124 -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 113 -
EleutherAI/pythia-31m
Text Generation • 55.7M • Updated • 3.58k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 97 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 95 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 89 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 85 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 96 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 98 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 95 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 87 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 93 -
EleutherAI/pythia-70m
95.6M • Updated • 70.6k • 78 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 192 -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 157 -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 159 -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 86 -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 89 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 85 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 79 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 69 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 71 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 692k • 38 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 316 -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 326 -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 318 -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 99 • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 75 -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 75 -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 70 -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 78 -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 65 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 45k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 557 -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 427 -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 249 -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 338 -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 165 -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 202 • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 165 -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 249 -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 236
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 72Note 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 • 70Note 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 • 65Note 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 • 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-seed2
Text Generation • Updated • 66Note 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 • 59Note 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