PolyPythias
- Preview • Updated • 1.76k
EleutherAI/pile-preshuffled-seeds
Updated • 151 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 8.95k • 30 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 1.35k -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 1.2k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 999 -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 924 -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 1.09k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 1.16k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 1.13k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.13k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 1.11k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 1.79k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 959 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 958 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 938 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 925 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 932 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 939 -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 960 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 988 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 950 -
EleutherAI/pythia-70m
95.6M • Updated • 560k • 85 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 1.22k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 943 -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 813 -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 787 -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 785 -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 779 -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 776 -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 778 -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 802 -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 2.45M • 44 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 2.01k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 1.77k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 1.43k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 893 • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 926 -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 842 -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 838 -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 841 -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 850 -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 317k • 39 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 1.36k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 1.38k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 1.18k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 1.09k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 1.05k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 972 • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 837 -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 839 -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 814
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 89.4kNote 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-data-seed2
Text Generation • Updated • 52.6kNote 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 • 51.8kNote 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 • 55.9kNote 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-seed2
Text Generation • Updated • 54.3kNote 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 • 51.5kNote 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