Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 2570 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4738. Min Act: -0.0001

Data Index: 462877 (C4 (Web Text))

Max Activating Token Index: 581

Text #1

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4627. Min Act: -0.0001

Data Index: 385327 (C4 (Web Text))

Max Activating Token Index: 426

Text #2

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4046. Min Act: -0.0001

Data Index: 248166 (C4 (Web Text))

Max Activating Token Index: 262

Text #3

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3972. Min Act: -0.0001

Data Index: 1306977 (C4 (Web Text))

Max Activating Token Index: 273

Text #4

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4329. Min Act: -0.0001

Data Index: 796529 (C4 (Web Text))

Max Activating Token Index: 823

Text #5

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4289. Min Act: -0.0001

Data Index: 313751 (C4 (Web Text))

Max Activating Token Index: 313

Text #6

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4078. Min Act: -0.0001

Data Index: 596967 (C4 (Web Text))

Max Activating Token Index: 147

Text #7

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4044. Min Act: -0.0001

Data Index: 677590 (C4 (Web Text))

Max Activating Token Index: 360

Text #8

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3807. Min Act: -0.0001

Data Index: 1151122 (C4 (Web Text))

Max Activating Token Index: 245

Text #9

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4194. Min Act: -0.0001

Data Index: 288299 (C4 (Web Text))

Max Activating Token Index: 660

Text #10

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3886. Min Act: -0.0001

Data Index: 1319331 (C4 (Web Text))

Max Activating Token Index: 682

Text #11

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3539. Min Act: -0.0001

Data Index: 495298 (C4 (Web Text))

Max Activating Token Index: 819

Text #12

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4060. Min Act: -0.0001

Data Index: 261317 (C4 (Web Text))

Max Activating Token Index: 557

Text #13

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3963. Min Act: -0.0001

Data Index: 1023073 (C4 (Web Text))

Max Activating Token Index: 242

Text #14

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4118. Min Act: -0.0001

Data Index: 192546 (C4 (Web Text))

Max Activating Token Index: 127

Text #15

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3875. Min Act: -0.0001

Data Index: 43297 (C4 (Web Text))

Max Activating Token Index: 488

Text #16

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4097. Min Act: -0.0001

Data Index: 389297 (C4 (Web Text))

Max Activating Token Index: 517

Text #17

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.4063. Min Act: -0.0001

Data Index: 1223047 (C4 (Web Text))

Max Activating Token Index: 653

Text #18

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3895. Min Act: -0.0001

Data Index: 998972 (C4 (Web Text))

Max Activating Token Index: 1004

Text #19

Max Range: 0.4738. Min Range: -0.4738

Max Act: 0.3833. Min Act: -0.0001

Data Index: 931683 (C4 (Web Text))

Max Activating Token Index: 966