PolyPythias
-
Preview • Updated • 146
EleutherAI/pile-preshuffled-seeds
Updated • 116 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m
Text Generation • 39.2M • Updated • 67.3k • 25 -
EleutherAI/pythia-14m-seed1
Updated • 383 -
EleutherAI/pythia-14m-seed2
Updated • 339 -
EleutherAI/pythia-14m-seed3
Updated • 52 -
EleutherAI/pythia-14m-seed4
Updated • 49 -
EleutherAI/pythia-14m-seed5
Updated • 24 -
EleutherAI/pythia-14m-seed6
Updated • 37 -
EleutherAI/pythia-14m-seed7
Updated • 25 -
EleutherAI/pythia-14m-seed8
Updated • 24 -
EleutherAI/pythia-14m-seed9
Updated • 42 -
EleutherAI/pythia-31m
Text Generation • 55.7M • Updated • 5.12k • 5 -
EleutherAI/pythia-31m-seed1
Updated • 11 -
EleutherAI/pythia-31m-seed2
Updated • 9 -
EleutherAI/pythia-31m-seed3
Updated • 10 -
EleutherAI/pythia-31m-seed4
Updated • 10 -
EleutherAI/pythia-31m-seed5
Updated • 4 -
EleutherAI/pythia-31m-seed6
Updated • 3 -
EleutherAI/pythia-31m-seed7
Updated • 3 -
EleutherAI/pythia-31m-seed8
Updated • 3 -
EleutherAI/pythia-31m-seed9
Updated • 3 -
EleutherAI/pythia-70m
95.6M • Updated • 167k • 74 -
EleutherAI/pythia-70m-seed1
Updated • 623 -
EleutherAI/pythia-70m-seed2
Updated • 578 -
EleutherAI/pythia-70m-seed3
Updated • 298 -
EleutherAI/pythia-70m-seed4
Updated • 10 -
EleutherAI/pythia-70m-seed5
Updated • 12 -
EleutherAI/pythia-70m-seed6
Updated • 17 -
EleutherAI/pythia-70m-seed7
Updated • 3 -
EleutherAI/pythia-70m-seed8
Updated • 4 -
EleutherAI/pythia-70m-seed9
Updated • 7 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 87.8k • 36 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 843 -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 836 -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 785 -
EleutherAI/pythia-160m-seed4
Updated • 11 • 1 -
EleutherAI/pythia-160m-seed5
Updated • 5 -
EleutherAI/pythia-160m-seed6
Updated • 5 -
EleutherAI/pythia-160m-seed7
Updated • 5 -
EleutherAI/pythia-160m-seed8
Updated • 5 -
EleutherAI/pythia-160m-seed9
Updated • 5 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 41.1k • 36 -
EleutherAI/pythia-410m-seed1
Updated • 13 -
EleutherAI/pythia-410m-seed2
Updated • 10 -
EleutherAI/pythia-410m-seed3
Updated • 11 -
EleutherAI/pythia-410m-seed4
Updated • 12 -
EleutherAI/pythia-410m-seed5
Updated • 6 -
EleutherAI/pythia-410m-seed6
Updated • 7 -
EleutherAI/pythia-410m-seed7
Updated • 6 -
EleutherAI/pythia-410m-seed8
Updated • 7 -
EleutherAI/pythia-410m-seed9
Updated • 7
EleutherAI/pythia-160m-data-seed1
Updated • 5Note 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
Updated • 4Note 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
Updated • 5Note 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
Updated • 5Note 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
Updated • 5Note 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
Updated • 5Note 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