Hugging Face's logo

Datasets:
bgradowhite
/
cells-developmental

Dataset card Data Studio Files
xet
 Community
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    TypeError
Message:      Couldn't cast array of type
struct<scans: double, crossing_median_degrees: double, crossing_finite_fraction: double, flip_median_degrees: double, flip_at_crossing: double, agreement_inside: double, agreement_outside: double, edge_median_degrees_l2_attn: double, edge_over_crossing_median_l2_attn: double, edge_median_degrees_l2_mlp: double, edge_over_crossing_median_l2_mlp: double, edge_median_degrees_kl: double, edge_over_crossing_median_kl: double, edge_median_degrees_fisher_rao: double, edge_over_crossing_median_fisher_rao: double, edge_median_degrees_hellinger: double, edge_over_crossing_median_hellinger: double>
to
{'scans': Value('float64'), 'crossing_median_degrees': Value('float64'), 'crossing_finite_fraction': Value('float64'), 'flip_median_degrees': Value('float64'), 'flip_at_crossing': Value('float64'), 'agreement_inside': Value('float64'), 'agreement_outside': Value('float64'), 'edge_median_degrees_l2_hidden': Value('float64'), 'edge_over_crossing_median_l2_hidden': Value('float64'), 'edge_median_degrees_kl': Value('float64'), 'edge_over_crossing_median_kl': Value('float64'), 'edge_median_degrees_fisher_rao': Value('float64'), 'edge_over_crossing_median_fisher_rao': Value('float64'), 'edge_median_degrees_hellinger': Value('float64'), 'edge_over_crossing_median_hellinger': Value('float64')}
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2312, in cast_table_to_schema
                  cast_array_to_feature(
                  ~~~~~~~~~~~~~~~~~~~~~^
                      table[name] if name in table_column_names else pa.array([None] * len(table), type=schema.field(name).type),
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                      feature,
                      ^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 1861, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ~~~~^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2158, in cast_array_to_feature
                  raise TypeError(f"Couldn't cast array of type\n{_short_str(array.type)}\nto\n{_short_str(feature)}")
              TypeError: Couldn't cast array of type
              struct<scans: double, crossing_median_degrees: double, crossing_finite_fraction: double, flip_median_degrees: double, flip_at_crossing: double, agreement_inside: double, agreement_outside: double, edge_median_degrees_l2_attn: double, edge_over_crossing_median_l2_attn: double, edge_median_degrees_l2_mlp: double, edge_over_crossing_median_l2_mlp: double, edge_median_degrees_kl: double, edge_over_crossing_median_kl: double, edge_median_degrees_fisher_rao: double, edge_over_crossing_median_fisher_rao: double, edge_median_degrees_hellinger: double, edge_over_crossing_median_hellinger: double>
              to
              {'scans': Value('float64'), 'crossing_median_degrees': Value('float64'), 'crossing_finite_fraction': Value('float64'), 'flip_median_degrees': Value('float64'), 'flip_at_crossing': Value('float64'), 'agreement_inside': Value('float64'), 'agreement_outside': Value('float64'), 'edge_median_degrees_l2_hidden': Value('float64'), 'edge_over_crossing_median_l2_hidden': Value('float64'), 'edge_median_degrees_kl': Value('float64'), 'edge_over_crossing_median_kl': Value('float64'), 'edge_median_degrees_fisher_rao': Value('float64'), 'edge_over_crossing_median_fisher_rao': Value('float64'), 'edge_median_degrees_hellinger': Value('float64'), 'edge_over_crossing_median_hellinger': Value('float64')}
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

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.

kind
string
substrate
string
cut
string
family
string
centre
string
bank_source
string
step
int64
scans
int64
ensure_mean
bool
evaluation
dict
summary
dict
angles
list
base_indices
list
crossing_degrees
list
flip_degrees
list
edge_degrees
dict
settled
dict
responses
dict
predicted
list
nearest_answer
list
metadata
dict
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
covariance_gaussian
bank_mean
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":70.0,"crossing_finite_fraction":1.0,"flip_median_degrees":19(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[70.0,72.5,72.5,67.5,72.5,72.5,65.0,60.0,65.0,77.5,72.5,72.5,75.0,65.0,67.5,70.0,62.5,70.0,65.0,72.5(...TRUNCATED)
[5.5,15.75,20.25,90.0,25.25,18.0,2.5,50.0,15.75,35.0,17.5,25.5,5.5,27.75,18.75,3.25,47.5,42.5,95.0,1(...TRUNCATED)
{"l2_hidden":[44.438,35.622,39.749,42.436,37.194,41.756,39.468,40.546,41.1,38.734,37.889,37.953,41.7(...TRUNCATED)
{"l2_hidden":[11.68795,10.301044,11.834296,11.987387,10.8849,11.30327,11.069753,13.165532,11.748185,(...TRUNCATED)
{"l2_hidden":[[3e-6,0.036817,0.073635,0.110456,0.147254,0.183866,0.220527,0.257245,0.293977,0.330523(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
covariance_gaussian
origin
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":67.5,"crossing_finite_fraction":1.0,"flip_median_degrees":19(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[70.0,70.0,72.5,67.5,72.5,70.0,62.5,60.0,65.0,75.0,72.5,72.5,75.0,65.0,67.5,70.0,62.5,67.5,65.0,72.5(...TRUNCATED)
[5.0,15.0,19.5,90.0,23.5,18.75,2.75,47.5,13.5,35.0,19.0,24.25,5.0,28.5,16.75,3.25,50.0,42.5,95.0,102(...TRUNCATED)
{"l2_hidden":[44.333,35.782,39.501,42.7,37.046,41.559,39.16,40.475,40.966,38.65,37.632,37.566,41.759(...TRUNCATED)
{"l2_hidden":[11.778938,10.446963,11.868155,11.971982,10.938555,11.454577,11.092997,13.140404,11.818(...TRUNCATED)
{"l2_hidden":[[3e-6,0.037137,0.074277,0.11142,0.148573,0.185573,0.222589,0.259665,0.296623,0.33352,0(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
random_uniform
bank_mean
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":72.5,"crossing_finite_fraction":1.0,"flip_median_degrees":21(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[77.5,75.0,72.5,72.5,75.0,75.0,75.0,72.5,70.0,75.0,72.5,75.0,77.5,75.0,77.5,67.5,77.5,80.0,72.5,75.0(...TRUNCATED)
[60.0,21.25,10.5,52.5,62.5,40.0,50.0,13.75,6.25,35.0,4.25,16.25,2.75,15.25,null,5.0,14.25,32.5,57.5,(...TRUNCATED)
{"l2_hidden":[39.597,38.157,38.969,42.154,39.873,40.469,41.18,39.834,41.547,39.288,35.962,40.689,43.(...TRUNCATED)
{"l2_hidden":[11.436117,10.417449,12.02208,12.247047,10.879272,11.137468,11.932301,13.037206,11.9479(...TRUNCATED)
{"l2_hidden":[[3e-6,0.040888,0.081762,0.122623,0.163472,0.204306,0.245126,0.285954,0.326807,0.367674(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
random_uniform
origin
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":72.5,"crossing_finite_fraction":1.0,"flip_median_degrees":21(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[77.5,75.0,70.0,72.5,75.0,75.0,75.0,72.5,67.5,75.0,72.5,75.0,77.5,75.0,77.5,67.5,77.5,80.0,72.5,75.0(...TRUNCATED)
[60.0,21.0,10.5,52.5,65.0,40.0,50.0,13.75,6.0,35.0,4.25,16.0,2.75,16.0,null,5.0,13.5,32.5,60.0,57.5,(...TRUNCATED)
{"l2_hidden":[39.683,38.37,38.97,41.826,39.542,40.22,41.19,39.728,41.567,38.973,35.801,40.398,44.116(...TRUNCATED)
{"l2_hidden":[11.5449,10.616587,12.02626,12.246519,10.854327,11.290269,11.963795,12.968729,12.047379(...TRUNCATED)
{"l2_hidden":[[3e-6,0.041281,0.082548,0.123801,0.165041,0.206265,0.247475,0.288696,0.329941,0.371198(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
real_activation
bank_mean
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":46.25,"crossing_finite_fraction":1.0,"flip_median_degrees":2(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,3809,910,964,244,1874,3114,1684,2586,2345,3047,2603,2615,1555,540,1570,86,1228,1084,593,1931,2(...TRUNCATED)
[42.5,45.0,47.5,45.0,47.5,47.5,47.5,45.0,45.0,47.5,45.0,47.5,45.0,45.0,47.5,47.5,47.5,42.5,45.0,47.5(...TRUNCATED)
[23.25,16.5,0.75,82.5,13.25,25.25,52.5,26.25,40.0,4.25,22.5,25.0,82.5,28.75,47.5,3.75,27.75,13.5,1.2(...TRUNCATED)
{"l2_hidden":[40.766,41.532,38.576,37.299,37.941,40.547,40.606,37.07,38.22,44.12,37.46,39.167,42.372(...TRUNCATED)
{"l2_hidden":[11.686634,10.910674,11.508519,12.063122,11.782201,12.099343,11.786673,10.441582,11.893(...TRUNCATED)
{"l2_hidden":[[3e-6,0.039267,0.07853,0.117787,0.15704,0.196134,0.23447,0.272917,0.311402,0.349688,0.(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
real_activation
origin
modular-train[p=113, frac=0.3, split 1000][n=3831]
0
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":45.0,"crossing_finite_fraction":1.0,"flip_median_degrees":27(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,3809,910,964,244,1874,3114,1684,2586,2345,3047,2603,2615,1555,540,1570,86,1228,1084,593,1931,2(...TRUNCATED)
[42.5,45.0,47.5,45.0,47.5,45.0,45.0,45.0,45.0,45.0,45.0,47.5,45.0,45.0,47.5,45.0,47.5,42.5,45.0,47.5(...TRUNCATED)
[21.0,15.75,0.75,82.5,13.25,22.75,50.0,28.0,35.0,40.0,21.75,26.75,82.5,29.25,47.5,3.75,27.25,12.75,1(...TRUNCATED)
{"l2_hidden":[40.758,41.672,38.654,37.151,37.809,40.506,40.533,37.023,37.979,43.99,37.551,39.071,42.(...TRUNCATED)
{"l2_hidden":[11.76115,11.001796,11.479624,12.104588,11.717865,12.140283,11.857945,10.423818,11.8204(...TRUNCATED)
{"l2_hidden":[[3e-6,0.039682,0.07936,0.119034,0.158703,0.198264,0.237043,0.275933,0.314778,0.353513,(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
covariance_gaussian
bank_mean
modular-train[p=113, frac=0.3, split 1000][n=3831]
1
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":70.0,"crossing_finite_fraction":1.0,"flip_median_degrees":21(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[70.0,72.5,72.5,67.5,72.5,72.5,65.0,60.0,65.0,77.5,72.5,72.5,75.0,65.0,67.5,70.0,62.5,70.0,65.0,72.5(...TRUNCATED)
[45.0,10.25,16.25,0.5,42.5,3.25,4.0,11.25,45.0,37.5,16.75,18.75,9.75,19.0,4.25,5.0,55.0,37.5,92.5,67(...TRUNCATED)
{"l2_hidden":[44.349,35.904,39.921,42.371,37.198,41.658,39.508,40.445,41.16,38.682,38.23,38.104,41.7(...TRUNCATED)
{"l2_hidden":[11.660228,10.325474,11.813269,11.949429,10.896237,11.325493,11.055696,13.151982,11.746(...TRUNCATED)
{"l2_hidden":[[3e-6,0.037406,0.074519,0.1114,0.14829,0.185226,0.222185,0.259158,0.29614,0.332981,0.3(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
covariance_gaussian
origin
modular-train[p=113, frac=0.3, split 1000][n=3831]
1
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":67.5,"crossing_finite_fraction":1.0,"flip_median_degrees":24(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[70.0,70.0,72.5,67.5,72.5,70.0,62.5,60.0,65.0,75.0,72.5,72.5,75.0,65.0,67.5,70.0,62.5,67.5,65.0,72.5(...TRUNCATED)
[42.5,9.75,16.25,0.5,42.5,3.25,4.25,10.5,42.5,37.5,14.75,15.5,9.5,25.5,3.5,3.25,50.0,40.0,92.5,65.0,(...TRUNCATED)
{"l2_hidden":[44.267,36.076,39.69,42.645,37.042,41.441,39.192,40.39,41.061,38.588,37.986,37.712,41.7(...TRUNCATED)
{"l2_hidden":[11.752238,10.475136,11.84755,11.933055,10.94746,11.473208,11.080754,13.127712,11.81917(...TRUNCATED)
{"l2_hidden":[[3e-6,0.037756,0.075209,0.112367,0.149583,0.186847,0.224134,0.261436,0.29875,0.336051,(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
random_uniform
bank_mean
modular-train[p=113, frac=0.3, split 1000][n=3831]
1
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":72.5,"crossing_finite_fraction":1.0,"flip_median_degrees":29(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[77.5,75.0,72.5,72.5,75.0,75.0,75.0,72.5,70.0,75.0,72.5,75.0,77.5,75.0,77.5,67.5,77.5,80.0,72.5,75.0(...TRUNCATED)
[65.0,14.5,17.75,1.25,67.5,11.75,3.0,12.0,52.5,40.0,60.0,6.75,1.75,32.5,null,1.5,47.5,27.75,62.5,50.(...TRUNCATED)
{"l2_hidden":[39.73,38.345,39.052,42.188,40.049,40.335,41.122,39.95,41.622,39.152,36.048,40.851,43.7(...TRUNCATED)
{"l2_hidden":[11.430514,10.435637,12.003727,12.211873,10.89697,11.141315,11.898618,13.029439,11.9418(...TRUNCATED)
{"l2_hidden":[[3e-6,0.040864,0.081725,0.122476,0.163156,0.203825,0.244491,0.285152,0.325804,0.36646,(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
token_boundary
modular_mlp:p113_d128_w512:mlp_p113_wd0.3_w512_d128_frac0.3_seed0
embed
random_uniform
origin
modular-train[p=113, frac=0.3, split 1000][n=3831]
1
40
true
{"evaluated_at":0,"step":0,"train_loss":4.7577009201049805,"train_acc":0.008352910168468952,"test_lo(...TRUNCATED)
{"scans":40.0,"crossing_median_degrees":72.5,"crossing_finite_fraction":1.0,"flip_median_degrees":32(...TRUNCATED)
[0.0,0.25,0.5,0.75,1.0,1.25,1.5,1.75,2.0,2.25,2.5,2.75,3.0,3.25,3.5,3.75,4.0,4.25,4.5,4.75,5.0,5.25,(...TRUNCATED)
[2195,2457,964,2541,3114,1565,2345,1520,2615,584,1570,2982,1084,1366,2263,3116,1844,1652,1955,3583,2(...TRUNCATED)
[77.5,75.0,70.0,72.5,75.0,75.0,75.0,72.5,67.5,75.0,72.5,75.0,77.5,75.0,77.5,67.5,77.5,80.0,72.5,75.0(...TRUNCATED)
[67.5,14.25,17.5,1.25,72.5,12.25,3.0,12.0,52.5,40.0,60.0,6.75,1.75,32.5,null,1.5,47.5,27.25,62.5,50.(...TRUNCATED)
{"l2_hidden":[39.795,38.558,39.072,41.867,39.699,40.095,41.168,39.863,41.646,38.856,35.915,40.558,44(...TRUNCATED)
{"l2_hidden":[11.540837,10.637211,12.014489,12.214107,10.867072,11.299277,11.93564,12.968691,12.0470(...TRUNCATED)
{"l2_hidden":[[3e-6,0.041257,0.08251,0.123648,0.164718,0.205776,0.246831,0.28788,0.328919,0.369965,0(...TRUNCATED)
[[26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26,26(...TRUNCATED)
[[42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42,42(...TRUNCATED)
{"seed":0,"settled_beyond_degrees":85.0,"edge_threshold":0.5,"codebook_entries":113,"slots":2,"degen(...TRUNCATED)
End of preview.

bgradowhite/cells-developmental

Verbatim backup of /Users/brianna/Cells_Developmental/out from Brianna's machine, taken 2026-09-14. See PROVENANCE.md for what each directory is and which script wrote it, and MANIFEST.json for SHA-256 digests of every file and every archive member. Archives <dir>.tar.gz extract in place to <dir>/; where a directory name held colons the archive name has hyphens instead, and MANIFEST.json's extract_to gives the original path.

The following is docs/hf-checkpoints-card.md from the source repository at backup time.

Cells_Developmental training trajectories

The trained networks the Cells_Developmental pages are measured on: every trajectory that the repository's own scripts produced, one directory per artifact, laid out exactly as it sits under out/ in a working copy. Copy a directory into out/ and the repository's loaders find it; inventory.json, config.json or log.json in each directory records the recipe, the seed and the checkpoint grid, and is the authority on what the directory holds.

Nothing here is a mirror of data that is public elsewhere. The MNIST and CIFAR archives, the Pythia checkpoint releases and the CIFAR-100 model-zoo trial are pinned by digest in the repository's configs/inputs/ and fetched from their sources; the modular-addition MLP trajectory of the curvature-steering runs is published separately. The measured parts that the pages are built from are not checkpoints and are not kept here.

Layout

directory network and recipe written by
mnist_task/ normalization-free residual tanh MLP on the pinned MNIST archives, five epoch checkpoints; the pinned artifact every other MNIST trajectory is compared against cells_developmental.exp1.mnist_task
mnist_dense/ the same network, data, recipe and seed, checkpointed on a log-spaced step grid scripts/train_mnist_dense.py
mnist_dense-b5/, mnist_dense-b6/ the same recipe and seed with five and six residual blocks scripts/train_mnist_dense.py --blocks
paper_mlp/plain/, paper_mlp/res/ the grokking paper's MNIST MLP (a thousand training points, MSE against one-hots, AdamW) and its minimal residual pair, on the 41-point checkpoint ladder continued to a million steps scripts/train_paper_mlp.py
paper_mlp/{plain,res}-h{4,5,6}/ the same pair at four, five and six hidden layers, on the 41-point ladder scripts/train_paper_mlp.py --hidden-layers
modular_transformer/<run>/ Nanda et al.'s one-layer transformer on modular addition, checkpointed on a logarithmic grid and at every evaluation where the test accuracy moved; latest.pt carries the optimiser state for resuming scripts/train_modular_transformer.py
vgg_cifar10/ VGG-19 on CIFAR-10 on Arora et al.'s recipe with the two departures the script's docstring records, checkpointed at 1, 2, 5, 10, 20, every fifteen epochs and the last scripts/train_vgg_cifar10.py

parts/<substrate>/ holds the measured parts behind a page, one JSON per panel named for what it measured, laid out flat as the repository's store reads them. The hub refuses a directory of more than ten thousand files, so a parts directory over that size is split here by panel kind (the second field of a part's name) with the overview left at the top; flatten it into out/parts/<substrate>/ before building.

Checkpoints ending in .npz are pickle-free NumPy archives; .pt files are torch.save state dictionaries and load with weights_only=True except for latest.pt.

Provenance

The bytes here are the ones measured on, not re-runs. Training is deterministic in the seed on one machine, but two machines running the same recipe do not produce the same bytes: a re-run of mnist_dense agreed with the original to about 1e-7 in every weight, while a re-run of paper_mlp/res had drifted by 4e-2 at step 1000. A trajectory retrained elsewhere is therefore a different object from the one the pages describe, and a directory is replaced here only by the copy that was measured on.

scripts/hf_upload_checkpoints.py in the repository compares the local artifacts with this tree by SHA-256 and sends only what is missing; a file that differs from its local counterpart is reported and left alone unless the artifact is named explicitly.
Downloads last month
382
Total file size:
10.2 GB