Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1543 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3989. Min Act: -0.0002

Data Index: 581984 (C4 (Web Text))

Max Activating Token Index: 579

Text #1

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3792. Min Act: -0.0002

Data Index: 429742 (C4 (Web Text))

Max Activating Token Index: 995

Text #2

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3369. Min Act: -0.0002

Data Index: 1143025 (C4 (Web Text))

Max Activating Token Index: 686

Text #3

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3260. Min Act: -0.0002

Data Index: 947062 (C4 (Web Text))

Max Activating Token Index: 273

Text #4

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3163. Min Act: -0.0003

Data Index: 1506658 (Python Code)

Max Activating Token Index: 230

Text #5

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3161. Min Act: -0.0002

Data Index: 485345 (C4 (Web Text))

Max Activating Token Index: 801

Text #6

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3237. Min Act: -0.0002

Data Index: 1472777 (Python Code)

Max Activating Token Index: 662

Text #7

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3160. Min Act: -0.0002

Data Index: 641151 (C4 (Web Text))

Max Activating Token Index: 785

Text #8

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3171. Min Act: -0.0002

Data Index: 989283 (C4 (Web Text))

Max Activating Token Index: 595

Text #9

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3156. Min Act: -0.0002

Data Index: 53936 (C4 (Web Text))

Max Activating Token Index: 161

Text #10

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3038. Min Act: -0.0003

Data Index: 596921 (C4 (Web Text))

Max Activating Token Index: 516

Text #11

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3104. Min Act: -0.0003

Data Index: 439016 (C4 (Web Text))

Max Activating Token Index: 902

Text #12

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3207. Min Act: -0.0002

Data Index: 809239 (C4 (Web Text))

Max Activating Token Index: 535

Text #13

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3097. Min Act: -0.0003

Data Index: 1506520 (Python Code)

Max Activating Token Index: 459

Text #14

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3191. Min Act: -0.0002

Data Index: 1236154 (C4 (Web Text))

Max Activating Token Index: 876

Text #15

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.2976. Min Act: -0.0003

Data Index: 1231373 (C4 (Web Text))

Max Activating Token Index: 481

Text #16

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3035. Min Act: -0.0003

Data Index: 1042646 (C4 (Web Text))

Max Activating Token Index: 488

Text #17

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3024. Min Act: -0.0002

Data Index: 1174597 (C4 (Web Text))

Max Activating Token Index: 516

Text #18

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.2918. Min Act: -0.0003

Data Index: 28493 (C4 (Web Text))

Max Activating Token Index: 677

Text #19

Max Range: 0.3989. Min Range: -0.3989

Max Act: 0.3113. Min Act: -0.0003

Data Index: 1262326 (C4 (Web Text))

Max Activating Token Index: 826