PolyPythias
- Preview • Updated • 3.12k
EleutherAI/pile-preshuffled-seeds
Updated • 337 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 14.3k • 30 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 2.29k -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 1.72k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 1.47k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 1.07k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 1.54k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 1.63k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 1.43k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.2k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 1.36k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 711 • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 835 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 839 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 829 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 753 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 723 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 798 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 714 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 699 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 730 -
EleutherAI/pythia-70m
95.6M • Updated • 475k • 85 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 2.99k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 1.96k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 1.54k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 1.01k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 953 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 992 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 865 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 905 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 910 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 3.5M • 45 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 44.7k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 41.5k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 33.9k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 17.1k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 17.3k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 17.7k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 7.1k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 7.05k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 6.18k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 205k • 40 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 14.3k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 16.7k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 5.76k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 5.05k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 4.95k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 4.81k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 4.86k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 5.65k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 6.22k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 67.1kNote 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 • 15kNote 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 • 10.4kNote 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 • 27kNote 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 • 14.9kNote 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 • 7.99kNote 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