Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 256 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3500. Min Act: -0.0000

Data Index: 552477 (C4 (Web Text))

Max Activating Token Index: 107

Text #1

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.4053. Min Act: -0.0000

Data Index: 467799 (C4 (Web Text))

Max Activating Token Index: 518

Text #2

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3935. Min Act: -0.0000

Data Index: 1337037 (C4 (Web Text))

Max Activating Token Index: 386

Text #3

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3891. Min Act: -0.0000

Data Index: 304998 (C4 (Web Text))

Max Activating Token Index: 541

Text #4

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.0678. Min Act: -0.0000

Data Index: 271486 (C4 (Web Text))

Max Activating Token Index: 107

Text #5

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.2278. Min Act: -0.0000

Data Index: 1491919 (Python Code)

Max Activating Token Index: 508

Text #6

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.0243. Min Act: -0.0000

Data Index: 96276 (C4 (Web Text))

Max Activating Token Index: 119

Text #7

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.1922. Min Act: -0.0000

Data Index: 819115 (C4 (Web Text))

Max Activating Token Index: 578

Text #8

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.2126. Min Act: -0.0000

Data Index: 1142101 (C4 (Web Text))

Max Activating Token Index: 230

Text #9

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3653. Min Act: -0.0000

Data Index: 285314 (C4 (Web Text))

Max Activating Token Index: 603

Text #10

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3550. Min Act: -0.0000

Data Index: 1398126 (Python Code)

Max Activating Token Index: 163

Text #11

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.1759. Min Act: -0.0000

Data Index: 1044639 (C4 (Web Text))

Max Activating Token Index: 26

Text #12

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.3135. Min Act: -0.0000

Data Index: 273003 (C4 (Web Text))

Max Activating Token Index: 869

Text #13

Max Range: 4.4053. Min Range: -4.4053

Max Act: 3.6330. Min Act: -0.0000

Data Index: 1574311 (Python Code)

Max Activating Token Index: 118

Text #14

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.1286. Min Act: -0.0000

Data Index: 936430 (C4 (Web Text))

Max Activating Token Index: 126

Text #15

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.1545. Min Act: -0.0000

Data Index: 1432282 (Python Code)

Max Activating Token Index: 711

Text #16

Max Range: 4.4053. Min Range: -4.4053

Max Act: 3.9482. Min Act: -0.0000

Data Index: 584365 (C4 (Web Text))

Max Activating Token Index: 60

Text #17

Max Range: 4.4053. Min Range: -4.4053

Max Act: 3.7801. Min Act: -0.0000

Data Index: 1608739 (Python Code)

Max Activating Token Index: 289

Text #18

Max Range: 4.4053. Min Range: -4.4053

Max Act: 3.7392. Min Act: -0.0000

Data Index: 1188334 (C4 (Web Text))

Max Activating Token Index: 850

Text #19

Max Range: 4.4053. Min Range: -4.4053

Max Act: 4.0849. Min Act: -0.0000

Data Index: 98653 (C4 (Web Text))

Max Activating Token Index: 1019