Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 2500 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.8454. Min Act: -0.0001

Data Index: 388717 (C4 (Web Text))

Max Activating Token Index: 108

Text #1

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7922. Min Act: -0.0001

Data Index: 468486 (C4 (Web Text))

Max Activating Token Index: 705

Text #2

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7801. Min Act: -0.0001

Data Index: 1086411 (C4 (Web Text))

Max Activating Token Index: 34

Text #3

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7630. Min Act: -0.0001

Data Index: 944805 (C4 (Web Text))

Max Activating Token Index: 33

Text #4

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.8143. Min Act: -0.0001

Data Index: 15369 (C4 (Web Text))

Max Activating Token Index: 24

Text #5

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.8044. Min Act: -0.0001

Data Index: 620592 (C4 (Web Text))

Max Activating Token Index: 168

Text #6

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7493. Min Act: -0.0001

Data Index: 1348995 (C4 (Web Text))

Max Activating Token Index: 752

Text #7

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7623. Min Act: -0.0001

Data Index: 957419 (C4 (Web Text))

Max Activating Token Index: 434

Text #8

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7382. Min Act: -0.0001

Data Index: 294653 (C4 (Web Text))

Max Activating Token Index: 46

Text #9

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7354. Min Act: -0.0001

Data Index: 1228894 (C4 (Web Text))

Max Activating Token Index: 51

Text #10

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7225. Min Act: -0.0001

Data Index: 871368 (C4 (Web Text))

Max Activating Token Index: 939

Text #11

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7184. Min Act: -0.0001

Data Index: 631233 (C4 (Web Text))

Max Activating Token Index: 428

Text #12

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7659. Min Act: -0.0001

Data Index: 521862 (C4 (Web Text))

Max Activating Token Index: 810

Text #13

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.6834. Min Act: -0.0001

Data Index: 230332 (C4 (Web Text))

Max Activating Token Index: 438

Text #14

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.6960. Min Act: -0.0001

Data Index: 434686 (C4 (Web Text))

Max Activating Token Index: 95

Text #15

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7141. Min Act: -0.0001

Data Index: 307168 (C4 (Web Text))

Max Activating Token Index: 621

Text #16

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.6656. Min Act: -0.0001

Data Index: 550044 (C4 (Web Text))

Max Activating Token Index: 111

Text #17

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.6847. Min Act: -0.0001

Data Index: 609681 (C4 (Web Text))

Max Activating Token Index: 776

Text #18

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.7014. Min Act: -0.0001

Data Index: 965169 (C4 (Web Text))

Max Activating Token Index: 630

Text #19

Max Range: 0.8454. Min Range: -0.8454

Max Act: 0.6753. Min Act: -0.0001

Data Index: 394136 (C4 (Web Text))

Max Activating Token Index: 167