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

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

Neuron 4180 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4836. Min Act: -0.0000

Data Index: 1307358 (C4 (Web Text))

Max Activating Token Index: 189

Text #1

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4853. Min Act: -0.0001

Data Index: 1502020 (Python Code)

Max Activating Token Index: 806

Text #2

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4597. Min Act: -0.0000

Data Index: 1482896 (Python Code)

Max Activating Token Index: 146

Text #3

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4350. Min Act: -0.0000

Data Index: 540768 (C4 (Web Text))

Max Activating Token Index: 446

Text #4

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4135. Min Act: -0.0001

Data Index: 1506930 (Python Code)

Max Activating Token Index: 875

Text #5

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4318. Min Act: -0.0000

Data Index: 643277 (C4 (Web Text))

Max Activating Token Index: 912

Text #6

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4155. Min Act: -0.0000

Data Index: 1359471 (C4 (Web Text))

Max Activating Token Index: 775

Text #7

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4100. Min Act: -0.0000

Data Index: 422559 (C4 (Web Text))

Max Activating Token Index: 46

Text #8

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.3702. Min Act: -0.0001

Data Index: 1635425 (Python Code)

Max Activating Token Index: 406

Text #9

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4198. Min Act: -0.0001

Data Index: 915688 (C4 (Web Text))

Max Activating Token Index: 21

Text #10

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4567. Min Act: -0.0001

Data Index: 1567958 (Python Code)

Max Activating Token Index: 986

Text #11

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4182. Min Act: -0.0001

Data Index: 78198 (C4 (Web Text))

Max Activating Token Index: 681

Text #12

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.3990. Min Act: -0.0000

Data Index: 1225377 (C4 (Web Text))

Max Activating Token Index: 8

Text #13

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4049. Min Act: -0.0000

Data Index: 289358 (C4 (Web Text))

Max Activating Token Index: 639

Text #14

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4369. Min Act: -0.0001

Data Index: 1602677 (Python Code)

Max Activating Token Index: 866

Text #15

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.3820. Min Act: -0.0001

Data Index: 1414641 (Python Code)

Max Activating Token Index: 545

Text #16

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4046. Min Act: -0.0000

Data Index: 1496502 (Python Code)

Max Activating Token Index: 321

Text #17

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.4037. Min Act: -0.0001

Data Index: 1003354 (C4 (Web Text))

Max Activating Token Index: 476

Text #18

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.3846. Min Act: -0.0001

Data Index: 1542193 (Python Code)

Max Activating Token Index: 889

Text #19

Max Range: 0.4853. Min Range: -0.4853

Max Act: 0.3964. Min Act: -0.0001

Data Index: 1270967 (C4 (Web Text))

Max Activating Token Index: 442