PolyPythias
- Preview • Updated • 83
EleutherAI/pile-preshuffled-seeds
Updated • 230 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 13.3k • 29 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 1.93k -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 1.61k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 1.57k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 1.54k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 1.79k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 1.74k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 1.73k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.76k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 1.74k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 1.49k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 1.48k -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 1.44k -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 1.44k -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 1.45k -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 1.47k -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 1.46k -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 1.48k -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 1.43k -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 1.49k -
EleutherAI/pythia-70m
95.6M • Updated • 554k • 81 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 4.46k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 4.51k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 4.41k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 4.41k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 4.19k -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 4.18k -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 3.93k -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 3.94k -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 3.91k -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 2.99M • 42 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 6.76k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 5.5k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 5.91k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 3.64k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 3.28k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 3.19k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 3.23k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 3.22k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 3.21k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 113k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 3.07k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 3.37k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 3.7k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 2.93k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 2.7k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 2.67k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 2.74k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 2.48k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 2.6k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 1.16kNote 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 • 1.14kNote 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 • 1.14kNote 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 • 1.16kNote 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 • 1.13kNote 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 • 1.14kNote 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