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

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

Neuron 634 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: 0.5164. Min Range: -0.5164

Max Act: 0.5164. Min Act: -0.0001

Data Index: 809866 (C4 (Web Text))

Max Activating Token Index: 731

Text #1

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3459. Min Act: -0.0001

Data Index: 469165 (C4 (Web Text))

Max Activating Token Index: 192

Text #2

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3184. Min Act: -0.0001

Data Index: 852282 (C4 (Web Text))

Max Activating Token Index: 730

Text #3

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3106. Min Act: -0.0001

Data Index: 679970 (C4 (Web Text))

Max Activating Token Index: 434

Text #4

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3220. Min Act: -0.0001

Data Index: 1482546 (Python Code)

Max Activating Token Index: 893

Text #5

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2855. Min Act: -0.0001

Data Index: 825235 (C4 (Web Text))

Max Activating Token Index: 164

Text #6

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2928. Min Act: -0.0001

Data Index: 94886 (C4 (Web Text))

Max Activating Token Index: 850

Text #7

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3033. Min Act: -0.0001

Data Index: 1007640 (C4 (Web Text))

Max Activating Token Index: 448

Text #8

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2977. Min Act: -0.0001

Data Index: 254722 (C4 (Web Text))

Max Activating Token Index: 855

Text #9

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2509. Min Act: -0.0001

Data Index: 310102 (C4 (Web Text))

Max Activating Token Index: 558

Text #10

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2699. Min Act: -0.0001

Data Index: 511080 (C4 (Web Text))

Max Activating Token Index: 30

Text #11

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2420. Min Act: -0.0001

Data Index: 541860 (C4 (Web Text))

Max Activating Token Index: 251

Text #12

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2840. Min Act: -0.0001

Data Index: 1632623 (Python Code)

Max Activating Token Index: 993

Text #13

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2689. Min Act: -0.0001

Data Index: 1513394 (Python Code)

Max Activating Token Index: 486

Text #14

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2598. Min Act: -0.0001

Data Index: 592223 (C4 (Web Text))

Max Activating Token Index: 366

Text #15

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2841. Min Act: -0.0001

Data Index: 1500565 (Python Code)

Max Activating Token Index: 204

Text #16

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2635. Min Act: -0.0001

Data Index: 1415197 (Python Code)

Max Activating Token Index: 152

Text #17

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2802. Min Act: -0.0001

Data Index: 1590822 (Python Code)

Max Activating Token Index: 975

Text #18

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.3168. Min Act: -0.0000

Data Index: 1302539 (C4 (Web Text))

Max Activating Token Index: 611

Text #19

Max Range: 0.5164. Min Range: -0.5164

Max Act: 0.2592. Min Act: -0.0001

Data Index: 827383 (C4 (Web Text))

Max Activating Token Index: 31