PolyPythias
- Preview • Updated • 1.05k
EleutherAI/pile-preshuffled-seeds
Updated • 1.03k • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 5.44k • 32 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 2.95k • 1 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 2.57k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 2.08k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 1.94k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 2.12k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 2.01k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 2.01k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.83k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 2.03k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 1.8k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 2.24k -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 2.02k -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 1.92k -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 1.83k -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 1.84k -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 1.91k -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 1.69k -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 1.71k -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 1.74k -
EleutherAI/pythia-70m
95.6M • Updated • 241k • 88 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 6.35k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 4.51k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 3.61k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 3.02k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 2.76k -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 2.95k -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 2.18k -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 1.98k -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 2.19k -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 3.49M • 66 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 38.3k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 16.8k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 15.5k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 3.22k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 3.05k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 2.25k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 2.37k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 2.27k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 2.44k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 93.4k • 41 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 4.34k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 3.43k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 3.48k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 3.29k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 2.69k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 2.32k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 2.36k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 2.04k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 2.26k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 2.22kNote 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 • 1.99kNote 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 • 1.89kNote 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 • 2.16kNote 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 • 2.03kNote 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 • 1.42kNote 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
EleutherAI/pythia-160m-weight-seed1-corrected
Text Generation • Updated • 1.03kNote Corrected weight-seed 1: independent initialization with the original Pythia data order.
EleutherAI/pythia-160m-weight-seed2-corrected
Text Generation • Updated • 851Note Corrected weight-seed 1: independent initialization with the original Pythia data order.