PolyPythias
- Preview • Updated • 74
EleutherAI/pile-preshuffled-seeds
Updated • 203 • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 15.3k • 29 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 2.17k -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 1.95k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 1.85k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 1.6k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 1.96k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 1.7k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 1.77k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.81k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 1.87k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 2.06k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 1.76k -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 1.65k -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 1.65k -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 1.76k -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 1.72k -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 1.73k -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 1.65k -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 1.62k -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 1.61k -
EleutherAI/pythia-70m
95.6M • Updated • 656k • 81 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 4.1k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 4.29k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 4.22k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 4.32k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 4.03k -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 4.27k -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 4.06k -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 3.91k -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 3.98k -
EleutherAI/pythia-160m
Text Generation • Updated • 3.03M • 42 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 7k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 6.26k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 6.29k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 3.82k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 3.6k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 3.28k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 3.33k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 3.29k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 3.73k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 147k • 36 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 3.8k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 4.26k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 3.55k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 3.2k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 3.5k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 3.2k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 3.28k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 3.17k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 3.51k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 1.13kNote 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 • 1.13kNote 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 • 1.12kNote 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 • 1.13kNote 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-seed2
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").
EleutherAI/pythia-160m-weight-seed3
Text Generation • Updated • 1.12kNote 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