PolyPythias
- Preview • Updated • 1.11k
EleutherAI/pile-preshuffled-seeds
Updated • 1.05k • 1Note Training data information for each seed.
-
EleutherAI/pythia-14m-deduped
Text Generation • 39.2M • Updated • 4.93k • 32 -
EleutherAI/pythia-14m-seed1
Text Generation • Updated • 3.47k • 1 -
EleutherAI/pythia-14m-seed2
Text Generation • Updated • 3.04k -
EleutherAI/pythia-14m-seed3
Text Generation • Updated • 2.48k -
EleutherAI/pythia-14m-seed4
Text Generation • Updated • 2.33k -
EleutherAI/pythia-14m-seed5
Text Generation • Updated • 2.34k -
EleutherAI/pythia-14m-seed6
Text Generation • Updated • 2.03k -
EleutherAI/pythia-14m-seed7
Text Generation • Updated • 2.04k -
EleutherAI/pythia-14m-seed8
Text Generation • Updated • 1.84k -
EleutherAI/pythia-14m-seed9
Text Generation • Updated • 2.05k -
EleutherAI/pythia-31m-deduped
Text Generation • 55.7M • Updated • 1.86k • 5 -
EleutherAI/pythia-31m-seed1
Text Generation • Updated • 2.69k -
EleutherAI/pythia-31m-seed2
Text Generation • Updated • 2.32k -
EleutherAI/pythia-31m-seed3
Text Generation • Updated • 2.15k -
EleutherAI/pythia-31m-seed4
Text Generation • Updated • 1.98k -
EleutherAI/pythia-31m-seed5
Text Generation • Updated • 1.83k -
EleutherAI/pythia-31m-seed6
Text Generation • Updated • 1.89k -
EleutherAI/pythia-31m-seed7
Text Generation • Updated • 1.67k -
EleutherAI/pythia-31m-seed8
Text Generation • Updated • 1.7k -
EleutherAI/pythia-31m-seed9
Text Generation • Updated • 1.73k -
EleutherAI/pythia-70m
95.6M • Updated • 244k • 88 -
EleutherAI/pythia-70m-seed1
Text Generation • Updated • 6.51k -
EleutherAI/pythia-70m-seed2
Text Generation • Updated • 4.81k -
EleutherAI/pythia-70m-seed3
Text Generation • Updated • 3.71k -
EleutherAI/pythia-70m-seed4
Text Generation • Updated • 3.08k -
EleutherAI/pythia-70m-seed5
Text Generation • Updated • 2.69k -
EleutherAI/pythia-70m-seed6
Text Generation • Updated • 2.81k -
EleutherAI/pythia-70m-seed7
Text Generation • Updated • 2.14k -
EleutherAI/pythia-70m-seed8
Text Generation • Updated • 1.94k -
EleutherAI/pythia-70m-seed9
Text Generation • Updated • 2.14k -
EleutherAI/pythia-160m
Text Generation • 0.2B • Updated • 3.5M • 66 -
EleutherAI/pythia-160m-seed1
Text Generation • 0.2B • Updated • 35.5k -
EleutherAI/pythia-160m-seed2
Text Generation • 0.2B • Updated • 13.7k -
EleutherAI/pythia-160m-seed3
Text Generation • 0.2B • Updated • 12k -
EleutherAI/pythia-160m-seed4
Text Generation • Updated • 3.29k • 1 -
EleutherAI/pythia-160m-seed5
Text Generation • Updated • 2.99k -
EleutherAI/pythia-160m-seed6
Text Generation • Updated • 2.2k -
EleutherAI/pythia-160m-seed7
Text Generation • Updated • 2.34k -
EleutherAI/pythia-160m-seed8
Text Generation • Updated • 2.18k -
EleutherAI/pythia-160m-seed9
Text Generation • Updated • 2.56k -
EleutherAI/pythia-410m
Text Generation • 0.5B • Updated • 94.8k • 41 -
EleutherAI/pythia-410m-seed1
Text Generation • Updated • 4.5k -
EleutherAI/pythia-410m-seed2
Text Generation • Updated • 3.47k -
EleutherAI/pythia-410m-seed3
Text Generation • Updated • 3.45k -
EleutherAI/pythia-410m-seed4
Text Generation • Updated • 3.23k -
EleutherAI/pythia-410m-seed5
Text Generation • Updated • 2.66k -
EleutherAI/pythia-410m-seed6
Text Generation • Updated • 2.29k • 1 -
EleutherAI/pythia-410m-seed7
Text Generation • Updated • 2.35k -
EleutherAI/pythia-410m-seed8
Text Generation • Updated • 2k -
EleutherAI/pythia-410m-seed9
Text Generation • Updated • 2.25k
EleutherAI/pythia-160m-data-seed1
Text Generation • Updated • 2.28kNote 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 • 2.04kNote 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 • 1.93kNote 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 • 2.18kNote 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 • 2.05kNote 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 • 1.43kNote 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 • 1.06kNote Corrected weight-seed 1: independent initialization with the original Pythia data order.
EleutherAI/pythia-160m-weight-seed2-corrected
Text Generation • Updated • 880Note Corrected weight-seed 1: independent initialization with the original Pythia data order.