Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 3867 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2889. Min Act: 0.0000

Data Index: 1239515 (C4 (Web Text))

Max Activating Token Index: 157

Text #1

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2780. Min Act: -0.0000

Data Index: 193814 (C4 (Web Text))

Max Activating Token Index: 772

Text #2

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2417. Min Act: -0.0001

Data Index: 959486 (C4 (Web Text))

Max Activating Token Index: 507

Text #3

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2258. Min Act: -0.0001

Data Index: 453988 (C4 (Web Text))

Max Activating Token Index: 588

Text #4

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2272. Min Act: -0.0000

Data Index: 898712 (C4 (Web Text))

Max Activating Token Index: 583

Text #5

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2164. Min Act: -0.0000

Data Index: 558113 (C4 (Web Text))

Max Activating Token Index: 794

Text #6

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2075. Min Act: -0.0001

Data Index: 431588 (C4 (Web Text))

Max Activating Token Index: 96

Text #7

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2152. Min Act: -0.0000

Data Index: 653541 (C4 (Web Text))

Max Activating Token Index: 461

Text #8

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2119. Min Act: -0.0000

Data Index: 321858 (C4 (Web Text))

Max Activating Token Index: 879

Text #9

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2170. Min Act: -0.0000

Data Index: 559243 (C4 (Web Text))

Max Activating Token Index: 254

Text #10

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2073. Min Act: -0.0000

Data Index: 807111 (C4 (Web Text))

Max Activating Token Index: 110

Text #11

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2031. Min Act: -0.0001

Data Index: 845326 (C4 (Web Text))

Max Activating Token Index: 918

Text #12

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2034. Min Act: -0.0000

Data Index: 730021 (C4 (Web Text))

Max Activating Token Index: 421

Text #13

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2028. Min Act: -0.0001

Data Index: 1359455 (C4 (Web Text))

Max Activating Token Index: 483

Text #14

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.1958. Min Act: -0.0001

Data Index: 938156 (C4 (Web Text))

Max Activating Token Index: 324

Text #15

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2019. Min Act: -0.0001

Data Index: 250878 (C4 (Web Text))

Max Activating Token Index: 935

Text #16

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2048. Min Act: -0.0001

Data Index: 1462618 (Python Code)

Max Activating Token Index: 817

Text #17

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.1925. Min Act: -0.0000

Data Index: 1314764 (C4 (Web Text))

Max Activating Token Index: 221

Text #18

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2045. Min Act: -0.0001

Data Index: 223897 (C4 (Web Text))

Max Activating Token Index: 851

Text #19

Max Range: 0.2889. Min Range: -0.2889

Max Act: 0.2037. Min Act: -0.0000

Data Index: 1133555 (C4 (Web Text))

Max Activating Token Index: 623