PolyPythias
- Preview • Updated • 797
EleutherAI/pile-preshuffled-seeds
Updated • 337 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 34.2k • 31 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 1.53k -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 1.31k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 1.14k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 1.02k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 1.21k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 2.36k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 1.1k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.06k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 1.04k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 2.46k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 797 -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 850 -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 680 -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 651 -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 660 -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 1.25k -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 612 -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 590 -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 625 -
EleutherAI/pythia-70m
95.6M • Updated • 502k • 85 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 4.42k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 2.82k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 2.11k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 1.54k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 1.49k -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 1.76k -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 1.42k -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 1.46k -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 1.44k -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 3.4M • 45 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 11.4k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 8.35k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 21.7k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 2.36k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 2.23k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 2.6k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 1.95k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 1.87k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 1.85k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 215k • 40 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 17.7k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 5.49k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 3.91k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 2.53k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 2.16k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 1.66k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 1.81k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 1.61k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 1.6k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 11.2kNote 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 • 4.08kNote 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 • 4.03kNote 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 • 5.15kNote 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 • 4.11kNote 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 • 3.97kNote 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