Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

Dataset: 80% C4 (Web Text) and 20% Python Code

Neuron 72 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0070. Min Act: -0.0006

Data Index: 1278121 (C4 (Web Text))

Max Activating Token Index: 399

Text #1

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0065. Min Act: -0.0006

Data Index: 402014 (C4 (Web Text))

Max Activating Token Index: 709

Text #2

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0059. Min Act: -0.0006

Data Index: 501952 (C4 (Web Text))

Max Activating Token Index: 193

Text #3

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0062. Min Act: -0.0006

Data Index: 1243226 (C4 (Web Text))

Max Activating Token Index: 361

Text #4

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0058. Min Act: -0.0006

Data Index: 1103775 (C4 (Web Text))

Max Activating Token Index: 387

Text #5

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0054. Min Act: -0.0006

Data Index: 341584 (C4 (Web Text))

Max Activating Token Index: 508

Text #6

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0057. Min Act: -0.0006

Data Index: 181696 (C4 (Web Text))

Max Activating Token Index: 423

Text #7

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0058. Min Act: -0.0006

Data Index: 31366 (C4 (Web Text))

Max Activating Token Index: 66

Text #8

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0056. Min Act: -0.0006

Data Index: 86585 (C4 (Web Text))

Max Activating Token Index: 478

Text #9

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0056. Min Act: -0.0006

Data Index: 824908 (C4 (Web Text))

Max Activating Token Index: 690

Text #10

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0055. Min Act: -0.0006

Data Index: 123664 (C4 (Web Text))

Max Activating Token Index: 136

Text #11

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0057. Min Act: -0.0006

Data Index: 620607 (C4 (Web Text))

Max Activating Token Index: 282

Text #12

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0053. Min Act: -0.0006

Data Index: 340250 (C4 (Web Text))

Max Activating Token Index: 930

Text #13

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0059. Min Act: -0.0006

Data Index: 326527 (C4 (Web Text))

Max Activating Token Index: 193

Text #14

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0051. Min Act: -0.0006

Data Index: 1342617 (C4 (Web Text))

Max Activating Token Index: 631

Text #15

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0056. Min Act: -0.0006

Data Index: 399381 (C4 (Web Text))

Max Activating Token Index: 820

Text #16

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0053. Min Act: -0.0006

Data Index: 490428 (C4 (Web Text))

Max Activating Token Index: 86

Text #17

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0050. Min Act: -0.0006

Data Index: 309370 (C4 (Web Text))

Max Activating Token Index: 1007

Text #18

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0050. Min Act: -0.0006

Data Index: 1139994 (C4 (Web Text))

Max Activating Token Index: 782

Text #19

Max Range: 0.0070. Min Range: -0.0070

Max Act: 0.0051. Min Act: -0.0006

Data Index: 1108804 (C4 (Web Text))

Max Activating Token Index: 1002