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

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

Neuron 3445 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.2148. Min Act: -0.0000

Data Index: 1557571 (Python Code)

Max Activating Token Index: 516

Text #1

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1602. Min Act: -0.0000

Data Index: 712454 (C4 (Web Text))

Max Activating Token Index: 639

Text #2

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1486. Min Act: -0.0000

Data Index: 371876 (C4 (Web Text))

Max Activating Token Index: 576

Text #3

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1524. Min Act: -0.0000

Data Index: 1554465 (Python Code)

Max Activating Token Index: 220

Text #4

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1250. Min Act: -0.0000

Data Index: 1313409 (C4 (Web Text))

Max Activating Token Index: 361

Text #5

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1323. Min Act: -0.0000

Data Index: 1391981 (Python Code)

Max Activating Token Index: 281

Text #6

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1138. Min Act: -0.0000

Data Index: 213619 (C4 (Web Text))

Max Activating Token Index: 389

Text #7

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1140. Min Act: -0.0000

Data Index: 1270531 (C4 (Web Text))

Max Activating Token Index: 958

Text #8

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1121. Min Act: -0.0000

Data Index: 514985 (C4 (Web Text))

Max Activating Token Index: 393

Text #9

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1092. Min Act: -0.0000

Data Index: 1586398 (Python Code)

Max Activating Token Index: 946

Text #10

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1067. Min Act: -0.0000

Data Index: 88466 (C4 (Web Text))

Max Activating Token Index: 640

Text #11

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1046. Min Act: -0.0000

Data Index: 1150378 (C4 (Web Text))

Max Activating Token Index: 650

Text #12

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0972. Min Act: -0.0000

Data Index: 1112722 (C4 (Web Text))

Max Activating Token Index: 664

Text #13

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0906. Min Act: -0.0000

Data Index: 1373950 (Python Code)

Max Activating Token Index: 988

Text #14

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0958. Min Act: -0.0000

Data Index: 1415130 (Python Code)

Max Activating Token Index: 605

Text #15

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0998. Min Act: -0.0000

Data Index: 1097955 (C4 (Web Text))

Max Activating Token Index: 483

Text #16

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0950. Min Act: -0.0000

Data Index: 1029859 (C4 (Web Text))

Max Activating Token Index: 1013

Text #17

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0970. Min Act: -0.0000

Data Index: 1005635 (C4 (Web Text))

Max Activating Token Index: 586

Text #18

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.1009. Min Act: -0.0000

Data Index: 1166435 (C4 (Web Text))

Max Activating Token Index: 936

Text #19

Max Range: 0.2148. Min Range: -0.2148

Max Act: 0.0968. Min Act: -0.0000

Data Index: 334820 (C4 (Web Text))

Max Activating Token Index: 816