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

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

Neuron 3666 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7738. Min Act: -0.0001

Data Index: 812768 (C4 (Web Text))

Max Activating Token Index: 873

Text #1

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7393. Min Act: -0.0001

Data Index: 872517 (C4 (Web Text))

Max Activating Token Index: 657

Text #2

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7427. Min Act: -0.0001

Data Index: 687440 (C4 (Web Text))

Max Activating Token Index: 731

Text #3

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7486. Min Act: -0.0001

Data Index: 1069160 (C4 (Web Text))

Max Activating Token Index: 40

Text #4

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7448. Min Act: -0.0001

Data Index: 863606 (C4 (Web Text))

Max Activating Token Index: 128

Text #5

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7141. Min Act: -0.0001

Data Index: 1303094 (C4 (Web Text))

Max Activating Token Index: 369

Text #6

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7058. Min Act: -0.0001

Data Index: 919725 (C4 (Web Text))

Max Activating Token Index: 873

Text #7

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7026. Min Act: -0.0001

Data Index: 220275 (C4 (Web Text))

Max Activating Token Index: 59

Text #8

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7214. Min Act: -0.0001

Data Index: 402125 (C4 (Web Text))

Max Activating Token Index: 683

Text #9

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7112. Min Act: -0.0001

Data Index: 610844 (C4 (Web Text))

Max Activating Token Index: 89

Text #10

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7018. Min Act: -0.0001

Data Index: 878003 (C4 (Web Text))

Max Activating Token Index: 45

Text #11

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6843. Min Act: -0.0001

Data Index: 678959 (C4 (Web Text))

Max Activating Token Index: 310

Text #12

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6741. Min Act: -0.0001

Data Index: 395508 (C4 (Web Text))

Max Activating Token Index: 290

Text #13

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6622. Min Act: -0.0001

Data Index: 470092 (C4 (Web Text))

Max Activating Token Index: 913

Text #14

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7148. Min Act: -0.0001

Data Index: 787956 (C4 (Web Text))

Max Activating Token Index: 527

Text #15

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6953. Min Act: -0.0001

Data Index: 1128103 (C4 (Web Text))

Max Activating Token Index: 471

Text #16

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.7193. Min Act: -0.0001

Data Index: 1180187 (C4 (Web Text))

Max Activating Token Index: 277

Text #17

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6868. Min Act: -0.0001

Data Index: 33911 (C4 (Web Text))

Max Activating Token Index: 1004

Text #18

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6808. Min Act: -0.0001

Data Index: 1232538 (C4 (Web Text))

Max Activating Token Index: 122

Text #19

Max Range: 0.7738. Min Range: -0.7738

Max Act: 0.6365. Min Act: -0.0001

Data Index: 1213514 (C4 (Web Text))

Max Activating Token Index: 799