Dataset Preview
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 77 new columns ({'weight_grad_norm/transformer.h.6.ln_1.weight', 'weight_grad_norm/transformer.h.5.ln_1.weight', 'empirical_L0_frac/transformer.h.6.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.4.mlp.c_proj.bias', 'empirical_L0_frac/transformer.h.7.attn.c_attn.weight', 'weight_grad_norm/transformer.h.4.ln_1.weight', 'weight_grad_norm/transformer.h.5.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.5.mlp.c_proj.bias', 'weight_grad_norm/transformer.h.6.mlp.c_fc.weight', 'empirical_L0_frac/transformer.h.6.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.5.attn.c_attn.bias', 'num_alive_neurons/c_fc/layer_5', 'weight_grad_norm/transformer.h.5.attn.c_attn.weight', 'empirical_L0_frac/transformer.h.6.attn.c_proj.weight', 'weight_grad_norm/transformer.h.6.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.4.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.4.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.5.mlp.c_proj.bias', 'num_alive_neurons/c_fc/layer_7', 'weight_grad_norm/transformer.h.4.attn.c_proj.weight', 'weight_grad_norm/transformer.h.6.ln_2.weight', 'empirical_L0_frac/transformer.h.6.mlp.c_proj.weight', 'empirical_L0_frac/transformer.h.4.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.6.attn.c_attn.weight', 'weight_grad_norm/transformer.h.5.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.6.mlp.c_proj.bias', 'empirical_L0_frac/transformer.h.7.mlp.c_proj.bias', 'empiric
...
 'empirical_L0_frac/transformer.h.7.attn.c_proj.bias', 'weight_grad_norm/transformer.h.5.ln_2.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_proj.weight', 'empirical_L0_frac/transformer.h.5.attn.c_attn.weight', 'empirical_L0_frac/transformer.h.4.attn.c_proj.weight', 'weight_grad_norm/transformer.h.5.mlp.c_fc.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.4.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.6.mlp.c_fc.bias', 'empirical_L0_frac/transformer.h.5.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.6.attn.c_attn.bias', 'empirical_L0_frac/transformer.h.7.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.5.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_proj.bias', 'weight_grad_norm/bigram_table', 'weight_grad_norm/transformer.h.6.attn.c_attn.weight', 'weight_grad_norm/transformer.h.5.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.5.attn.c_attn.bias', 'weight_grad_norm/transformer.h.7.attn.c_attn.bias', 'empirical_L0_frac/transformer.h.6.mlp.c_proj.bias', 'weight_grad_norm/transformer.h.4.attn.c_attn.weight', 'weight_grad_norm/transformer.h.7.attn.c_attn.weight', 'weight_grad_norm/transformer.h.7.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.7.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.7.attn.c_proj.bias', 'num_alive_neurons/c_fc/layer_4', 'empirical_L0_frac/transformer.h.7.attn.c_attn.bias', 'weight_grad_norm/transformer.h.4.ln_2.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_fc.weight'}) and 6 missing columns ({'dead_qk/layer_1', 'test_xent_bridged', 'test_kl_bridged', 'dead_qk/layer_0', 'dead_qk/layer_2', 'dead_qk/layer_3'}).

This happened while the json dataset builder was generating data using

hf://datasets/michaelwaves/sparse-circuits/train_curves/csp_sweep1_16x_7.4Mnonzero_afrac0.125/progress.json (at revision 15dadd2c2824ca52bcbcb7ae58533d1998bd454e)

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1831, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              xent: double
              test_kl: double
              test_xent: double
              grad_norm: int64
              weight_sparsity: double
              L0_as_frac_of_orig_params: double
              L0: double
              L0_non_LN: double
              L0_non_embed: double
              L0_non_embed_as_frac_of_orig_params: double
              param_norm: double
              aux_invertable: int64
              elapsed_tokens: int64
              grad_scale: double
              lr: double
              pfrac: int64
              step: int64
              did_clip_grad_norm: int64
              tokens_per_second: double
              num_alive_neurons/c_fc/layer_0: int64
              num_alive_neurons/c_fc/layer_1: int64
              num_alive_neurons/c_fc/layer_2: int64
              num_alive_neurons/c_fc/layer_3: int64
              num_alive_neurons/c_fc/layer_4: int64
              num_alive_neurons/c_fc/layer_5: int64
              num_alive_neurons/c_fc/layer_6: int64
              num_alive_neurons/c_fc/layer_7: int64
              empirical_L0_frac/transformer.wte.weight: double
              empirical_L0_frac/transformer.wpe.weight: double
              empirical_L0_frac/transformer.h.0.attn.c_attn.weight: double
              empirical_L0_frac/transformer.h.0.attn.c_attn.bias: double
              empirical_L0_frac/transformer.h.0.attn.c_proj.weight: double
              empirical_L0_frac/transformer.h.0.attn.c_proj.bias: double
              empirical_L0_frac/transformer.h.0.mlp.c_fc.weight: double
              empirical_L0_frac/transformer.h.0.mlp.c_fc.bias: double
              empirical_L0_frac/transformer.h.0.mlp.c_proj.weight: double
              empirical_L0_frac/transformer.h.0.mlp.c_proj.bias: double
              empirical_L0_frac/transformer.h.1.attn.c_attn.weight: double
              empirical_L0_frac/transformer.h.1.attn.c_attn.bias: double
              empirical_L0_frac/transformer.h.1.attn.c_proj.weight: double
              empirical_L0_frac/transformer.h.1.attn.c_proj.bias: double
              
              ...
              ad_norm/transformer.h.5.ln_2.weight: double
              weight_grad_norm/transformer.h.5.mlp.c_fc.weight: double
              weight_grad_norm/transformer.h.5.mlp.c_fc.bias: double
              weight_grad_norm/transformer.h.5.mlp.c_proj.weight: double
              weight_grad_norm/transformer.h.5.mlp.c_proj.bias: double
              weight_grad_norm/transformer.h.6.ln_1.weight: double
              weight_grad_norm/transformer.h.6.attn.c_attn.weight: double
              weight_grad_norm/transformer.h.6.attn.c_attn.bias: double
              weight_grad_norm/transformer.h.6.attn.c_proj.weight: double
              weight_grad_norm/transformer.h.6.attn.c_proj.bias: double
              weight_grad_norm/transformer.h.6.ln_2.weight: double
              weight_grad_norm/transformer.h.6.mlp.c_fc.weight: double
              weight_grad_norm/transformer.h.6.mlp.c_fc.bias: double
              weight_grad_norm/transformer.h.6.mlp.c_proj.weight: double
              weight_grad_norm/transformer.h.6.mlp.c_proj.bias: double
              weight_grad_norm/transformer.h.7.ln_1.weight: double
              weight_grad_norm/transformer.h.7.attn.c_attn.weight: double
              weight_grad_norm/transformer.h.7.attn.c_attn.bias: double
              weight_grad_norm/transformer.h.7.attn.c_proj.weight: double
              weight_grad_norm/transformer.h.7.attn.c_proj.bias: double
              weight_grad_norm/transformer.h.7.ln_2.weight: double
              weight_grad_norm/transformer.h.7.mlp.c_fc.weight: double
              weight_grad_norm/transformer.h.7.mlp.c_fc.bias: double
              weight_grad_norm/transformer.h.7.mlp.c_proj.weight: double
              weight_grad_norm/transformer.h.7.mlp.c_proj.bias: double
              weight_grad_norm/transformer.ln_f.weight: double
              weight_grad_norm/lm_head.weight: double
              to
              {'xent': Value('float64'), 'test_kl': Value('float64'), 'test_xent': Value('float64'), 'grad_norm': Value('int64'), 'weight_sparsity': Value('float64'), 'L0_as_frac_of_orig_params': Value('float64'), 'L0': Value('float64'), 'L0_non_LN': Value('float64'), 'L0_non_embed': Value('float64'), 'L0_non_embed_as_frac_of_orig_params': Value('float64'), 'param_norm': Value('float64'), 'aux_invertable': Value('float64'), 'elapsed_tokens': Value('int64'), 'grad_scale': Value('float64'), 'lr': Value('float64'), 'pfrac': Value('float64'), 'step': Value('int64'), 'did_clip_grad_norm': Value('int64'), 'tokens_per_second': Value('float64'), 'num_alive_neurons/c_fc/layer_0': Value('int64'), 'num_alive_neurons/c_fc/layer_1': Value('int64'), 'num_alive_neurons/c_fc/layer_2': Value('int64'), 'num_alive_neurons/c_fc/layer_3': Value('int64'), 'test_kl_bridged': Value('float64'), 'test_xent_bridged': Value('float64'), 'dead_qk/layer_0': Value('int64'), 'dead_qk/layer_1': Value('int64'), 'dead_qk/layer_2': Value('int64'), 'dead_qk/layer_3': Value('int64'), 'empirical_L0_frac/transformer.wte.weight': Value('float64'), 'empirical_L0_frac/transformer.wpe.weight': Value('float64'), 'empirical_L0_frac/transformer.h.0.attn.c_attn.weight': Value('float64'), 'empirical_L0_frac/transformer.h.0.attn.c_attn.bias': Value('float64'), 'empirical_L0_frac/transformer.h.0.attn.c_proj.weight': Value('float64'), 'empirical_L0_frac/transformer.h.0.attn.c_proj.bias': Value('float64'), 'empirical_L0_frac/transformer.h.0.m
              ...
              orm/transformer.h.2.ln_1.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.attn.c_attn.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.attn.c_attn.bias': Value('float64'), 'weight_grad_norm/transformer.h.2.attn.c_proj.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.attn.c_proj.bias': Value('float64'), 'weight_grad_norm/transformer.h.2.ln_2.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.mlp.c_fc.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.mlp.c_fc.bias': Value('float64'), 'weight_grad_norm/transformer.h.2.mlp.c_proj.weight': Value('float64'), 'weight_grad_norm/transformer.h.2.mlp.c_proj.bias': Value('float64'), 'weight_grad_norm/transformer.h.3.ln_1.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.attn.c_attn.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.attn.c_attn.bias': Value('float64'), 'weight_grad_norm/transformer.h.3.attn.c_proj.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.attn.c_proj.bias': Value('float64'), 'weight_grad_norm/transformer.h.3.ln_2.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.mlp.c_fc.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.mlp.c_fc.bias': Value('float64'), 'weight_grad_norm/transformer.h.3.mlp.c_proj.weight': Value('float64'), 'weight_grad_norm/transformer.h.3.mlp.c_proj.bias': Value('float64'), 'weight_grad_norm/transformer.ln_f.weight': Value('float64'), 'weight_grad_norm/lm_head.weight': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1450, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 993, in stream_convert_to_parquet
                  builder._prepare_split(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1833, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 77 new columns ({'weight_grad_norm/transformer.h.6.ln_1.weight', 'weight_grad_norm/transformer.h.5.ln_1.weight', 'empirical_L0_frac/transformer.h.6.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.4.mlp.c_proj.bias', 'empirical_L0_frac/transformer.h.7.attn.c_attn.weight', 'weight_grad_norm/transformer.h.4.ln_1.weight', 'weight_grad_norm/transformer.h.5.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.5.mlp.c_proj.bias', 'weight_grad_norm/transformer.h.6.mlp.c_fc.weight', 'empirical_L0_frac/transformer.h.6.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.5.attn.c_attn.bias', 'num_alive_neurons/c_fc/layer_5', 'weight_grad_norm/transformer.h.5.attn.c_attn.weight', 'empirical_L0_frac/transformer.h.6.attn.c_proj.weight', 'weight_grad_norm/transformer.h.6.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.4.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.4.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.5.mlp.c_proj.bias', 'num_alive_neurons/c_fc/layer_7', 'weight_grad_norm/transformer.h.4.attn.c_proj.weight', 'weight_grad_norm/transformer.h.6.ln_2.weight', 'empirical_L0_frac/transformer.h.6.mlp.c_proj.weight', 'empirical_L0_frac/transformer.h.4.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.6.attn.c_attn.weight', 'weight_grad_norm/transformer.h.5.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.6.mlp.c_proj.bias', 'empirical_L0_frac/transformer.h.7.mlp.c_proj.bias', 'empiric
              ...
               'empirical_L0_frac/transformer.h.7.attn.c_proj.bias', 'weight_grad_norm/transformer.h.5.ln_2.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_proj.weight', 'empirical_L0_frac/transformer.h.5.attn.c_attn.weight', 'empirical_L0_frac/transformer.h.4.attn.c_proj.weight', 'weight_grad_norm/transformer.h.5.mlp.c_fc.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.4.attn.c_proj.bias', 'empirical_L0_frac/transformer.h.6.mlp.c_fc.bias', 'empirical_L0_frac/transformer.h.5.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.6.attn.c_attn.bias', 'empirical_L0_frac/transformer.h.7.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.5.mlp.c_proj.weight', 'weight_grad_norm/transformer.h.4.mlp.c_proj.bias', 'weight_grad_norm/bigram_table', 'weight_grad_norm/transformer.h.6.attn.c_attn.weight', 'weight_grad_norm/transformer.h.5.attn.c_proj.weight', 'empirical_L0_frac/transformer.h.5.attn.c_attn.bias', 'weight_grad_norm/transformer.h.7.attn.c_attn.bias', 'empirical_L0_frac/transformer.h.6.mlp.c_proj.bias', 'weight_grad_norm/transformer.h.4.attn.c_attn.weight', 'weight_grad_norm/transformer.h.7.attn.c_attn.weight', 'weight_grad_norm/transformer.h.7.mlp.c_fc.bias', 'weight_grad_norm/transformer.h.7.mlp.c_fc.weight', 'weight_grad_norm/transformer.h.7.attn.c_proj.bias', 'num_alive_neurons/c_fc/layer_4', 'empirical_L0_frac/transformer.h.7.attn.c_attn.bias', 'weight_grad_norm/transformer.h.4.ln_2.weight', 'empirical_L0_frac/transformer.h.7.mlp.c_fc.weight'}) and 6 missing columns ({'dead_qk/layer_1', 'test_xent_bridged', 'test_kl_bridged', 'dead_qk/layer_0', 'dead_qk/layer_2', 'dead_qk/layer_3'}).
              
              This happened while the json dataset builder was generating data using
              
              hf://datasets/michaelwaves/sparse-circuits/train_curves/csp_sweep1_16x_7.4Mnonzero_afrac0.125/progress.json (at revision 15dadd2c2824ca52bcbcb7ae58533d1998bd454e)
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

xent
float64
test_kl
float64
test_xent
float64
grad_norm
int64
weight_sparsity
float64
L0_as_frac_of_orig_params
float64
L0
float64
L0_non_LN
float64
L0_non_embed
float64
L0_non_embed_as_frac_of_orig_params
float64
param_norm
float64
aux_invertable
float64
elapsed_tokens
int64
grad_scale
float64
lr
float64
pfrac
float64
step
int64
did_clip_grad_norm
int64
tokens_per_second
float64
num_alive_neurons/c_fc/layer_0
int64
num_alive_neurons/c_fc/layer_1
int64
num_alive_neurons/c_fc/layer_2
int64
num_alive_neurons/c_fc/layer_3
int64
test_kl_bridged
float64
test_xent_bridged
float64
dead_qk/layer_0
int64
dead_qk/layer_1
int64
dead_qk/layer_2
int64
dead_qk/layer_3
int64
empirical_L0_frac/transformer.wte.weight
float64
empirical_L0_frac/transformer.wpe.weight
float64
empirical_L0_frac/transformer.h.0.attn.c_attn.weight
float64
empirical_L0_frac/transformer.h.0.attn.c_attn.bias
float64
empirical_L0_frac/transformer.h.0.attn.c_proj.weight
float64
empirical_L0_frac/transformer.h.0.attn.c_proj.bias
float64
empirical_L0_frac/transformer.h.0.mlp.c_fc.weight
float64
empirical_L0_frac/transformer.h.0.mlp.c_fc.bias
float64
empirical_L0_frac/transformer.h.0.mlp.c_proj.weight
float64
empirical_L0_frac/transformer.h.0.mlp.c_proj.bias
float64
empirical_L0_frac/transformer.h.1.attn.c_attn.weight
float64
empirical_L0_frac/transformer.h.1.attn.c_attn.bias
float64
empirical_L0_frac/transformer.h.1.attn.c_proj.weight
float64
empirical_L0_frac/transformer.h.1.attn.c_proj.bias
float64
empirical_L0_frac/transformer.h.1.mlp.c_fc.weight
float64
empirical_L0_frac/transformer.h.1.mlp.c_fc.bias
float64
empirical_L0_frac/transformer.h.1.mlp.c_proj.weight
float64
empirical_L0_frac/transformer.h.1.mlp.c_proj.bias
float64
empirical_L0_frac/transformer.h.2.attn.c_attn.weight
float64
empirical_L0_frac/transformer.h.2.attn.c_attn.bias
float64
empirical_L0_frac/transformer.h.2.attn.c_proj.weight
float64
empirical_L0_frac/transformer.h.2.attn.c_proj.bias
float64
empirical_L0_frac/transformer.h.2.mlp.c_fc.weight
float64
empirical_L0_frac/transformer.h.2.mlp.c_fc.bias
float64
empirical_L0_frac/transformer.h.2.mlp.c_proj.weight
float64
empirical_L0_frac/transformer.h.2.mlp.c_proj.bias
float64
empirical_L0_frac/transformer.h.3.attn.c_attn.weight
float64
empirical_L0_frac/transformer.h.3.attn.c_attn.bias
float64
empirical_L0_frac/transformer.h.3.attn.c_proj.weight
float64
empirical_L0_frac/transformer.h.3.attn.c_proj.bias
float64
empirical_L0_frac/transformer.h.3.mlp.c_fc.weight
float64
empirical_L0_frac/transformer.h.3.mlp.c_fc.bias
float64
empirical_L0_frac/transformer.h.3.mlp.c_proj.weight
float64
empirical_L0_frac/transformer.h.3.mlp.c_proj.bias
float64
empirical_L0_frac/lm_head.weight
float64
weight_grad_norm/transformer.wte.weight
float64
weight_grad_norm/transformer.wpe.weight
float64
weight_grad_norm/transformer.h.0.ln_1.weight
float64
weight_grad_norm/transformer.h.0.attn.c_attn.weight
float64
weight_grad_norm/transformer.h.0.attn.c_attn.bias
float64
weight_grad_norm/transformer.h.0.attn.c_proj.weight
float64
weight_grad_norm/transformer.h.0.attn.c_proj.bias
float64
weight_grad_norm/transformer.h.0.ln_2.weight
float64
weight_grad_norm/transformer.h.0.mlp.c_fc.weight
float64
weight_grad_norm/transformer.h.0.mlp.c_fc.bias
float64
weight_grad_norm/transformer.h.0.mlp.c_proj.weight
float64
weight_grad_norm/transformer.h.0.mlp.c_proj.bias
float64
weight_grad_norm/transformer.h.1.ln_1.weight
float64
weight_grad_norm/transformer.h.1.attn.c_attn.weight
float64
weight_grad_norm/transformer.h.1.attn.c_attn.bias
float64
weight_grad_norm/transformer.h.1.attn.c_proj.weight
float64
weight_grad_norm/transformer.h.1.attn.c_proj.bias
float64
weight_grad_norm/transformer.h.1.ln_2.weight
float64
weight_grad_norm/transformer.h.1.mlp.c_fc.weight
float64
weight_grad_norm/transformer.h.1.mlp.c_fc.bias
float64
weight_grad_norm/transformer.h.1.mlp.c_proj.weight
float64
weight_grad_norm/transformer.h.1.mlp.c_proj.bias
float64
weight_grad_norm/transformer.h.2.ln_1.weight
float64
weight_grad_norm/transformer.h.2.attn.c_attn.weight
float64
weight_grad_norm/transformer.h.2.attn.c_attn.bias
float64
weight_grad_norm/transformer.h.2.attn.c_proj.weight
float64
weight_grad_norm/transformer.h.2.attn.c_proj.bias
float64
weight_grad_norm/transformer.h.2.ln_2.weight
float64
weight_grad_norm/transformer.h.2.mlp.c_fc.weight
float64
weight_grad_norm/transformer.h.2.mlp.c_fc.bias
float64
weight_grad_norm/transformer.h.2.mlp.c_proj.weight
float64
weight_grad_norm/transformer.h.2.mlp.c_proj.bias
float64
weight_grad_norm/transformer.h.3.ln_1.weight
float64
weight_grad_norm/transformer.h.3.attn.c_attn.weight
float64
weight_grad_norm/transformer.h.3.attn.c_attn.bias
float64
weight_grad_norm/transformer.h.3.attn.c_proj.weight
float64
weight_grad_norm/transformer.h.3.attn.c_proj.bias
float64
weight_grad_norm/transformer.h.3.ln_2.weight
float64
weight_grad_norm/transformer.h.3.mlp.c_fc.weight
float64
weight_grad_norm/transformer.h.3.mlp.c_fc.bias
float64
weight_grad_norm/transformer.h.3.mlp.c_proj.weight
float64
weight_grad_norm/transformer.h.3.mlp.c_proj.bias
float64
weight_grad_norm/transformer.ln_f.weight
float64
weight_grad_norm/lm_head.weight
float64
1.371694
0
1.276191
0
0.989376
0.259892
2,251,264
2,228,224
1,835,008
0.5
610.702698
0.594859
31,992,053,760
8,192
0.000006
0.5
61,020
0
269,959.871277
8,192
8,192
8,192
8,192
0.203483
1.365613
1,546
1,711
1,610
1,299
0.0625
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.007813
0.0625
0.0625
1,994.947876
1,419.624512
42.788998
3,418.453125
76.63591
2,001.557251
45.179031
42.6329
4,028.899658
85.289574
3,970.382813
45.125263
42.638042
3,463.61499
76.2995
2,001.994141
45.133698
41.742344
4,005.307373
85.306297
3,873.67627
45.10429
41.697556
3,443.843506
75.443939
1,986.846558
45.098782
38.983082
3,926.290039
81.894409
3,875.323242
45.052746
42.817844
3,390.329102
74.335213
1,973.016968
45.038067
41.031883
3,912.85376
80.599854
3,904.460449
44.765808
43.085915
2,006.281128
7.599796
0
7.548134
0
0.000348
24.454414
211,830,784
211,812,352
205,520,896
56
271.653412
25.803831
0
65,536
0
64
0
0
0
8,192
8,192
8,192
8,192
6.313459
7.549939
0
0
0
0
1
1
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
2,042.441895
1,445.860107
44.536713
3,417.94751
78.21595
2,043.262329
45.25259
45.196499
4,006.147949
90.136925
4,088.913574
45.245182
44.620705
3,454.940186
77.868614
2,045.24939
45.247963
45.226604
4,031.810059
90.204514
4,088.154297
45.239128
44.14304
3,348.273926
77.44931
2,041.631104
45.24192
45.206886
3,970.539063
90.018944
4,080.6521
45.227757
43.025169
3,029.654785
76.186371
2,029.564209
45.224136
42.688358
3,560.497559
88.277634
4,063.51123
45.201862
45.206448
2,044.870728
6.098747
0
6.018192
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
271.395508
28.006756
10,485,760
65,536
0.000074
64
20
0
213,724.714137
8,192
8,192
8,192
8,192
4.868958
6.110586
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,024.374146
1,436.76123
44.634747
3,101.237549
77.239395
2,020.609863
45.229256
45.111858
3,913.091797
89.088036
4,046.999756
45.173038
43.59293
3,180.97583
74.38829
2,034.434082
45.18993
45.19162
3,989.652344
89.820747
4,043.778564
45.158539
43.672146
3,025.878662
74.698257
2,028.325684
45.19117
45.166584
3,899.591797
89.589149
4,028.159424
45.139484
44.003937
2,991.673096
74.756332
2,023.001587
45.173389
44.834373
3,523.896484
86.968803
4,022.80249
45.106232
45.147758
2,044.234741
5.70662
0
5.536369
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.967529
20.301104
20,971,520
65,536
0.000148
64
40
0
240,548.875229
8,192
8,192
8,192
8,192
4.218064
5.458321
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,994.526611
1,411.366699
44.119457
2,903.021973
77.060318
2,018.457642
45.211472
44.953495
3,979.353271
88.883781
4,054.255127
45.181942
44.423656
3,294.275879
74.72554
2,032.884766
45.170803
44.953133
3,892.654297
88.349258
3,986.199707
45.111355
44.110325
3,187.261719
73.550255
2,027.265503
45.13969
44.741943
3,697.235352
86.732254
3,868.342773
45.028236
42.864365
2,873.369873
69.685135
2,009.54834
45.028992
41.814026
2,983.696777
73.369652
3,880.479492
44.362804
45.007381
2,037.780396
5.532545
0
5.424904
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.557312
17.471241
31,457,280
65,536
0.000223
64
60
0
250,908.126209
8,192
8,192
8,192
8,192
3.81137
5.054487
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,990.942017
1,401.326416
44.705479
3,378.902832
77.881409
2,038.997803
45.23761
45.198154
4,046.885742
89.55011
4,062.207031
45.188034
44.344284
3,379.884521
75.551064
2,033.253052
45.183643
44.552868
3,866.267822
85.773163
3,979.109375
45.093315
43.767799
3,253.471924
73.656914
2,020.601074
45.129738
43.751408
3,628.754395
81.14782
3,796.562988
45.042522
41.091221
2,952.963135
68.854263
2,002.120483
45.024139
44.347645
3,735.443848
84.717499
3,955.677734
44.676514
45.066643
2,038.4552
5.094608
0
4.943391
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.305481
14.104076
41,943,040
65,536
0.000297
64
80
0
256,611.114932
8,192
8,192
8,192
8,192
3.355571
4.598867
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,965.004883
1,401.592285
44.713261
3,319.762695
77.704666
2,034.22522
45.231434
45.136143
4,022.622314
89.228767
4,059.048584
45.197479
44.223785
3,404.084473
75.816971
2,033.121948
45.195683
44.421791
3,870.101807
86.002388
4,018.85083
45.15007
42.723808
3,180.932617
72.384048
2,012.970581
45.137669
42.832153
3,567.246826
81.712166
3,808.503906
45.063221
40.785126
2,953.167725
69.659592
1,996.739136
45.07056
44.598991
3,861.677979
86.865356
4,016.754395
44.983814
45.154896
2,040.440186
4.746141
0
4.516998
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.002808
14.97901
52,428,800
65,536
0.000371
64
100
0
260,174.723843
8,192
8,192
8,192
8,192
2.809456
4.051325
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,953.960327
1,397.453491
44.352894
3,268.207275
77.60891
2,034.867676
45.235916
45.124626
4,042.440674
89.668152
4,055.394287
45.216038
44.257931
3,402.859131
75.942123
2,031.130737
45.208008
44.300045
3,839.019043
85.519547
4,019.004883
45.194153
44.06839
3,343.174316
75.537903
2,024.53479
45.202026
43.814575
3,786.081299
85.735489
3,921.450928
45.171505
44.501114
3,372.421875
76.172287
2,027.473022
45.195721
45.111279
4,034.821289
89.639641
4,052.7146
45.140812
45.154015
2,041.812744
4.382658
0
4.197106
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
269.786774
11.52207
62,914,560
65,536
0.000445
64
120
0
262,405.57835
8,192
8,192
8,192
8,192
2.432196
3.672437
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,943.880737
1,391.163086
43.353951
3,073.774414
76.84272
2,035.953125
45.223454
44.890465
4,002.887939
88.976418
4,048.109131
45.208481
44.357281
3,404.838379
76.092644
2,027.36084
45.210346
44.245064
3,796.461182
84.530975
4,005.958984
45.199272
44.093517
3,356.951904
76.035828
2,014.947754
45.19677
43.302837
3,601.930908
83.228111
3,904.180908
45.177204
43.992844
3,292.799316
75.310486
2,022.112915
45.193531
45.036919
4,014.365967
89.470764
4,045.775635
45.168247
45.157513
2,042.953613
4.157256
0
3.968408
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
269.721008
10.186149
73,400,320
65,536
0.000519
64
140
0
263,528.317851
8,192
8,192
8,192
8,192
2.105189
3.343932
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,949.997681
1,390.417969
43.413353
3,148.731689
77.009407
2,033.994507
45.22113
44.890182
3,985.474609
88.722229
4,034.059082
45.203938
44.531403
3,420.216309
76.578812
2,018.591919
45.207005
44.47155
3,747.083496
84.143539
3,984.544434
45.204979
44.055717
3,368.928223
75.935593
2,015.072144
45.199032
44.12875
3,831.817871
86.658104
3,919.103271
45.182602
44.317524
3,388.469971
76.444595
2,021.802979
45.20002
44.966862
3,987.887207
89.162735
4,033.185791
45.182808
45.150105
2,041.941895
4.14553
0
3.855193
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
269.841431
10.204352
83,886,080
65,536
0.000594
64
160
0
264,656.319895
8,192
8,192
8,192
8,192
1.847092
3.086919
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,982.688843
1,404.646484
44.369148
3,303.03833
77.802773
2,038.697876
45.23914
45.062649
4,034.096924
89.62738
4,052.11499
45.229252
44.626781
3,444.42627
76.763298
2,023.836182
45.222057
44.700764
3,856.671387
86.567978
4,007.808594
45.223038
44.086475
3,373.118896
76.17189
2,021.080322
45.215492
44.168781
3,757.440186
85.81002
3,969.983398
45.215351
44.401894
3,371.402588
76.402405
2,028.637573
45.218845
45.008472
4,003.691895
89.358414
4,033.064453
45.205132
45.168327
2,043.710449
3.931593
0
3.700436
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.100342
8.666575
94,371,840
65,536
0.000668
64
180
0
265,817.515167
8,192
8,192
8,192
8,192
1.679852
2.916628
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,977.75708
1,396.908813
44.365215
3,350.705811
77.792961
2,035.373535
45.23587
45.026825
4,017.502686
89.450615
4,038.096191
45.223675
44.533112
3,431.784424
76.789238
2,014.824951
45.216816
44.571026
3,824.945313
86.012909
3,986.557129
45.220295
44.331661
3,378.228516
76.490448
2,010.748657
45.209576
44.172115
3,780.362305
86.201302
3,932.653809
45.207973
44.936104
3,471.90918
77.604935
2,029.384521
45.221043
44.987503
4,001.912598
89.243553
4,017.787842
45.20047
45.158562
2,042.523315
3.772708
0
3.574709
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.499237
7.900064
104,857,600
65,536
0.000742
64
200
0
266,621.900123
8,192
8,192
8,192
8,192
1.578905
2.814312
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,976.03894
1,395.300049
44.313335
3,361.035645
77.72142
2,034.007324
45.233257
45.062134
3,999.512939
89.174538
4,022.727295
45.212288
44.182457
3,383.47583
75.935356
2,005.854736
45.206139
44.241222
3,773.335693
85.250587
3,963.743408
45.214199
44.975449
3,485.955078
77.747467
2,024.953369
45.224972
44.672905
3,860.925781
86.819389
3,929.117432
45.210255
44.421253
3,390.026855
76.485886
2,021.778564
45.207867
44.966011
3,991.077148
89.182457
4,018.359375
45.197422
45.152317
2,043.224976
3.632998
0
3.447376
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
270.973206
6.714726
115,343,360
65,536
0.000816
64
220
0
265,777.564476
8,192
8,192
8,192
8,192
1.499292
2.732976
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,973.661987
1,398.305176
44.357887
3,323.031494
77.645187
2,031.979004
45.229954
44.909496
3,986.919434
89.198776
4,027.865479
45.216019
44.52597
3,416.169189
76.470573
2,004.623901
45.210857
44.508389
3,811.934814
86.240044
3,963.161865
45.218002
44.310562
3,387.773926
76.40657
2,009.022217
45.208157
44.386112
3,784.291504
85.888077
3,917.313477
45.2085
44.901306
3,481.619385
77.683716
2,024.924072
45.221668
45.014095
3,975.718262
88.740852
4,000.212891
45.199039
45.151825
2,042.557617
3.382951
0
3.282897
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
271.503723
5.905764
125,829,120
65,536
0.00089
64
240
0
260,235.441335
8,192
8,192
8,192
8,192
1.434669
2.668562
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,984.384521
1,403.081177
44.523594
3,356.570068
77.687096
2,033.611694
45.231682
44.889694
3,970.364502
89.042297
4,021.224609
45.216702
44.528107
3,437.839111
76.941422
2,002.032104
45.212711
44.333702
3,743.568115
85.177422
3,950.19458
45.21302
44.513306
3,405.756348
76.71682
2,007.994019
45.209362
44.312672
3,782.957764
85.96109
3,905.260742
45.204365
44.732761
3,433.446289
77.206993
2,017.310791
45.212322
44.895863
3,957.097168
88.390831
3,989.552979
45.19363
45.1469
2,042.918945
3.288423
0
3.15651
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
272.111908
5.976633
136,314,880
65,536
0.000964
64
260
0
239,102.937429
8,192
8,192
8,192
8,192
1.400091
2.632823
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,999.109985
1,407.929932
44.719254
3,405.048828
77.791039
2,035.313599
45.235191
45.038849
4,005.349365
89.400665
4,044.837646
45.222744
44.426807
3,425.602539
76.708557
2,003.915649
45.213387
44.859077
3,880.963623
86.673523
3,973.288574
45.219536
44.560585
3,434.814697
77.030251
2,012.672119
45.215385
44.845615
3,940.984863
87.614021
3,949.706787
45.208912
45.001366
3,499.11377
77.875175
2,020.672729
45.222729
45.044144
4,007.729004
88.91729
4,015.262695
45.196301
45.159195
2,042.896118
3.206962
0
2.986931
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
272.769257
5.418612
146,800,640
65,536
0.001039
64
280
0
241,312.568348
8,192
8,192
8,192
8,192
1.320852
2.551461
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,990.045776
1,403.986206
44.67831
3,382.977051
77.565254
2,030.391479
45.22744
44.958878
3,983.324463
89.197571
4,026.377686
45.216522
44.598339
3,443.709229
76.879608
1,991.983887
45.206886
44.445274
3,795.392578
85.793098
3,934.55249
45.211533
44.487793
3,429.194336
77.11869
2,000.282715
45.20702
44.411484
3,850.766602
86.116875
3,873.680664
45.195965
44.702576
3,438.47876
77.226044
2,011.500488
45.201466
45.082737
3,993.719727
88.814896
3,975.368896
45.185181
45.118359
2,040.852051
3.061177
0
2.887115
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
273.389099
4.909681
157,286,400
65,536
0.001113
64
300
0
243,187.830758
8,192
8,192
8,192
8,192
1.300131
2.528109
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,996.364868
1,405.875488
44.87603
3,434.700928
77.590103
2,032.457642
45.227634
44.926941
3,967.894287
88.872574
4,036.104736
45.217491
44.620026
3,439.8479
76.888382
1,997.838623
45.208427
44.568966
3,864.932373
86.388298
3,944.916504
45.209484
44.51931
3,415.895264
76.564537
2,001.504395
45.20153
44.408924
3,841.245605
85.954163
3,885.658936
45.194798
44.383568
3,399.176025
76.412903
2,006.117798
45.194004
45.01498
3,984.093506
88.34391
3,986.854248
45.169418
45.130726
2,042.196289
2.947598
0
2.799381
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
274.064789
4.216142
167,772,160
65,536
0.001187
64
320
0
244,674.706155
8,192
8,192
8,192
8,192
1.292536
2.519907
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,991.436035
1,406.244751
44.867184
3,433.229736
77.651917
2,030.817383
45.229538
44.99519
3,993.063232
89.153618
4,041.993652
45.218189
44.540459
3,433.367676
76.776588
1,995.753662
45.204346
44.524078
3,844.953369
86.006592
3,957.445557
45.215839
44.680031
3,461.9104
77.238586
2,009.12085
45.211205
44.704376
3,902.932373
86.803085
3,925.201904
45.201496
44.805466
3,470.05957
77.391037
2,013.935791
45.204884
45.06855
4,007.065674
88.873039
4,004.919434
45.185398
45.150024
2,042.994019
3.000859
0
2.709356
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
274.779449
4.402226
178,257,920
65,536
0.001261
64
340
0
246,089.160294
8,192
8,192
8,192
8,192
1.213526
2.439985
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,990.609985
1,407.446533
44.935364
3,446.432373
77.825348
2,032.114502
45.234863
45.12801
4,038.263184
89.838112
4,034.5896
45.224792
44.697941
3,451.755859
77.122375
2,008.840332
45.219913
44.992245
3,992.849854
88.500984
3,975.477295
45.225067
44.905869
3,490.225098
77.656647
2,018.049316
45.221043
44.94812
3,991.211914
88.438629
3,910.025635
45.209122
44.913361
3,470.744873
77.439484
2,017.883911
45.211132
45.094086
4,012.364014
89.135277
3,962.010742
45.195602
45.163902
2,042.292114
2.743012
0
2.662057
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
275.505188
4.196299
188,743,680
65,536
0.001335
64
360
0
247,416.899297
8,192
8,192
8,192
8,192
1.196619
2.422268
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,000.154785
1,406.037964
44.960979
3,455.569824
77.622955
2,029.617798
45.230736
45.074715
4,013.252197
89.325027
4,034.20166
45.220554
44.725937
3,465.890381
77.306061
2,005.086426
45.215797
44.777691
3,908.763672
87.083641
3,954.101074
45.219662
44.729385
3,452.878174
77.263138
2,013.474121
45.2187
44.751541
3,909.052979
87.079445
3,903.493896
45.210438
44.992855
3,489.304443
77.743782
2,024.337769
45.216557
45.03204
3,993.471436
88.73819
3,979.350342
45.194454
45.155262
2,043.317017
2.935627
0
2.626852
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
276.309448
3.563018
199,229,440
65,536
0.00141
64
380
0
248,607.591339
8,192
8,192
8,192
8,192
1.161115
2.386936
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,986.538818
1,406.953369
44.96854
3,448.187988
77.775909
2,029.629028
45.236012
44.954514
3,991.970947
89.449539
4,028.50293
45.2286
44.797054
3,469.131348
77.370399
2,012.535034
45.222107
44.654076
3,873.610107
87.331116
3,959.132813
45.223503
44.832676
3,475.960449
77.567017
2,021.059326
45.226135
44.881199
3,910.285645
87.38015
3,897.246582
45.210629
45.076336
3,504.46167
77.919762
2,029.934448
45.225368
45.006542
4,004.596924
88.91803
3,972.26123
45.199593
45.153633
2,042.194824
2.660415
0
2.507989
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
277.142303
3.698636
209,715,200
65,536
0.001484
64
400
0
249,678.156955
8,192
8,192
8,192
8,192
1.129611
2.354918
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,994.184326
1,407.811035
44.793747
3,429.135742
77.594872
2,025.869629
45.229347
45.004276
3,975.853516
88.748474
4,005.952148
45.213799
44.75119
3,460.038086
77.304977
2,008.464844
45.216625
44.72929
3,872.258545
86.711449
3,922.81665
45.209381
44.533772
3,431.251953
77.064034
2,006.918457
45.21077
44.41222
3,802.080078
85.401733
3,844.48291
45.191273
44.688282
3,436.72583
77.0261
2,016.598633
45.204273
44.860172
3,952.006348
88.000237
3,930.860352
45.172604
45.111214
2,041.567749
2.769478
0
2.483879
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
278.117493
4.226847
220,200,960
65,536
0.001558
64
420
0
250,589.458957
8,192
8,192
8,192
8,192
1.093285
2.317593
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,996.036987
1,411.414673
44.751827
3,406.759277
77.537971
2,025.771973
45.230633
45.063877
4,001.771729
89.418694
4,016.380615
45.227211
44.938072
3,493.544189
77.628853
2,016.816528
45.22504
44.853024
3,885.219727
87.118683
3,940.37207
45.22197
44.843254
3,478.426514
77.562492
2,017.56958
45.221973
44.85918
3,885.420166
87.03743
3,873.047363
45.209278
44.86779
3,474.615234
77.439682
2,023.983154
45.215752
44.81337
3,955.335693
88.238045
3,922.816895
45.188538
45.122726
2,041.239258
2.612691
0
2.446553
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
279.209106
3.868529
230,686,720
65,536
0.001632
64
440
0
251,491.711326
8,192
8,192
8,192
8,192
1.066457
2.288887
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,989.515381
1,403.659058
44.968441
3,457.553955
77.774429
2,030.028198
45.232372
45.139275
4,034.000488
89.571259
4,023.698486
45.215588
44.829494
3,472.702148
77.521973
2,009.822876
45.220261
44.987411
3,981.010498
88.3564
3,912.218018
45.210464
44.840492
3,473.810547
77.5177
2,014.749634
45.218307
44.735374
3,931.927734
87.353409
3,881.867432
45.197811
44.817787
3,481.003906
77.591576
2,021.778564
45.214535
45.105049
4,012.301025
88.959221
3,967.248779
45.181606
45.144428
2,042.500122
2.630618
0
2.399622
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
280.261627
4.092134
241,172,480
65,536
0.001706
64
460
0
252,240.786649
8,192
8,192
8,192
8,192
1.043359
2.26375
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,998.661743
1,408.186646
44.950905
3,457.096191
77.540771
2,025.762573
45.223976
44.838924
3,967.864502
88.622276
4,024.515381
45.21373
44.916451
3,472.505615
77.477318
2,002.992432
45.215988
44.998093
3,957.244385
87.850784
3,907.601318
45.215031
44.788517
3,473.003906
77.463356
2,007.58667
45.214798
44.829086
3,899.875977
87.076981
3,899.912842
45.203453
44.85915
3,493.007813
77.721916
2,017.508423
45.212528
44.918774
3,986.417725
88.454155
3,987.551025
45.166973
45.157871
2,044.232544
2.538602
0
2.500082
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
281.551178
3.265527
251,658,240
65,536
0.001781
64
480
0
252,963.232662
8,192
8,192
8,192
8,192
1.039091
2.260105
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,008.236694
1,411.05542
45.068672
3,494.746094
78.112801
2,036.623413
45.244335
45.159237
4,035.511475
89.763466
4,041.180908
45.235382
44.993126
3,505.351074
77.9095
2,027.093628
45.236732
44.952827
3,976.319336
88.38176
3,958.091797
45.230846
45.055721
3,507.425049
77.950645
2,029.22583
45.235985
45.023918
3,912.449951
87.786926
3,952.237549
45.228046
45.181141
3,521.417725
78.067688
2,036.142944
45.235973
45.109417
4,039.502686
89.516487
4,003.661865
45.219723
45.212326
2,045.100586
2.528949
0
2.369133
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
282.805298
3.430302
262,144,000
65,536
0.001855
64
500
0
253,687.972676
8,192
8,192
8,192
8,192
0.98545
2.206796
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,002.567505
1,408.890381
45.005383
3,475.081543
77.914177
2,030.78064
45.236877
45.104706
4,017.535645
89.482887
4,024.710205
45.229645
44.994038
3,504.776367
77.882454
2,018.428833
45.230621
45.058552
3,948.174805
87.970032
3,955.111572
45.227169
45.097481
3,507.92749
77.925545
2,025.075806
45.230198
44.786488
3,936.344971
87.612083
3,870.808838
45.213509
45.061558
3,504.305908
77.827728
2,026.011963
45.221901
45.050316
4,004.692627
88.959671
3,941.344971
45.194191
45.157921
2,043.500854
2.62236
0
2.319871
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
284.214996
3.339241
272,629,760
65,536
0.001929
64
520
0
254,303.60887
8,192
8,192
8,192
8,192
0.94149
2.159912
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,992.931763
1,407.256714
44.922417
3,465.544189
77.93177
2,031.526123
45.236763
45.174019
4,045.464111
89.64962
4,007.259521
45.218109
44.917976
3,489.055664
77.711029
1,998.745361
45.219406
45.057117
4,008.825195
88.823303
3,907.482666
45.216953
44.929729
3,498.848145
77.812683
2,013.19104
45.222637
44.860043
3,931.248291
87.432747
3,863.592529
45.20657
44.991028
3,500.609619
77.79303
2,022.606323
45.216003
44.935749
3,968.21875
88.226448
3,920.237305
45.172741
45.173203
2,042.387573
2.540878
0
2.33543
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
285.789215
3.620048
283,115,520
65,536
0.002003
64
540
0
254,829.251205
8,192
8,192
8,192
8,192
0.924922
2.143721
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,001.269897
1,406.792969
44.976364
3,472.436523
77.859299
2,027.019165
45.231113
45.113335
4,026.301514
89.471054
4,013.638184
45.224648
45.150425
3,512.192139
77.989738
2,014.485229
45.230392
45.119308
3,985.038818
88.475342
3,927.46167
45.22422
45.124947
3,503.42749
77.903305
2,016.920166
45.226612
44.898445
3,969.034424
88.072021
3,849.730225
45.207172
45.120396
3,504.542236
77.839546
2,025.021606
45.219261
44.853077
3,972.868408
88.257523
3,921.938477
45.181305
45.133377
2,042.392334
2.622222
0
2.291544
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
287.494263
3.10755
293,601,280
65,536
0.002077
64
560
0
255,318.444708
8,192
8,192
8,192
8,192
0.893954
2.1119
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,002.685303
1,409.379517
45.069389
3,489.991943
77.962219
2,028.504395
45.231907
44.97797
4,012.666992
89.291664
4,028.987305
45.220764
45.048203
3,497.368896
77.832199
2,005.294312
45.223804
45.108994
3,982.526367
88.553352
3,905.811035
45.222607
44.935532
3,495.68042
77.80246
2,003.866211
45.218914
44.928802
3,915.160889
87.632179
3,845.235596
45.20845
45.042282
3,499.108154
77.791595
2,020.835815
45.214554
45.152821
4,030.941895
89.267456
3,962.96582
45.177708
45.173306
2,042.811646
2.344009
0
2.313026
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
289.219635
3.333708
304,087,040
65,536
0.002151
64
580
0
255,720.945216
8,192
8,192
8,192
8,192
0.889694
2.10679
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1,991.565308
1,406.252197
44.940853
3,458.116211
77.788467
2,027.431519
45.232082
45.077019
4,006.949951
89.414146
4,025.932129
45.229935
45.17989
3,514.951904
78.028534
2,017.02356
45.232811
45.106033
3,974.262451
88.322662
3,946.166992
45.228317
45.195457
3,521.765381
78.103065
2,024.209351
45.235004
45.007675
3,982.036377
88.405983
3,922.212891
45.218113
45.139309
3,525.162354
78.105614
2,026.944214
45.225491
45.114616
4,023.548096
89.205795
3,959.779053
45.192211
45.193558
2,043.390381
2.536633
0
2.280575
0
0
24.462925
211,904,512
211,812,352
205,520,896
56
291.144684
3.176618
314,572,800
65,536
0.002226
64
600
0
256,243.058821
8,192
8,192
8,192
8,192
0.864328
2.082856
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2,008.65686
1,407.285278
45.082706
3,491.775879
78.000854
2,028.374634
45.234703
45.125671
4,022.549561
89.473076
4,006.842041
45.220158
45.122219
3,505.006348
77.908951
2,006.106689
45.223877
45.08197
3,994.25415
88.595604
3,905.425049
45.219433
45.179813
3,509.094238
77.97551
2,017.565308
45.229168
44.95369
3,901.039307
86.999878
3,834.873779
45.204964
45.188953
3,515.250977
77.978111
2,023.312378
45.221592
45.105286
4,015.1604
89.000938
3,878.013428
45.170361
45.174049
2,043.362793
End of preview.
YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

##About

Publicly available artifacts from OpenAI's paper Weight-sparse transformers have interpretable circuits

https://cdn.openai.com/pdf/41df8f28-d4ef-43e9-aed2-823f9393e470/circuit-sparsity-paper.pdf

Downloads last month
17