Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 41 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5848. Min Act: -0.0000

Data Index: 597940 (C4 (Web Text))

Max Activating Token Index: 22

Text #1

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5617. Min Act: -0.0000

Data Index: 933568 (C4 (Web Text))

Max Activating Token Index: 883

Text #2

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5124. Min Act: -0.0000

Data Index: 899157 (C4 (Web Text))

Max Activating Token Index: 269

Text #3

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5815. Min Act: -0.0000

Data Index: 1378521 (Python Code)

Max Activating Token Index: 285

Text #4

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5003. Min Act: -0.0000

Data Index: 1555227 (Python Code)

Max Activating Token Index: 512

Text #5

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.5143. Min Act: -0.0000

Data Index: 512879 (C4 (Web Text))

Max Activating Token Index: 220

Text #6

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4855. Min Act: -0.0000

Data Index: 412911 (C4 (Web Text))

Max Activating Token Index: 941

Text #7

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4908. Min Act: -0.0000

Data Index: 136193 (C4 (Web Text))

Max Activating Token Index: 362

Text #8

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4488. Min Act: -0.0000

Data Index: 405091 (C4 (Web Text))

Max Activating Token Index: 163

Text #9

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4473. Min Act: -0.0000

Data Index: 491628 (C4 (Web Text))

Max Activating Token Index: 653

Text #10

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4739. Min Act: -0.0000

Data Index: 717385 (C4 (Web Text))

Max Activating Token Index: 441

Text #11

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4547. Min Act: -0.0000

Data Index: 482767 (C4 (Web Text))

Max Activating Token Index: 14

Text #12

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4667. Min Act: -0.0000

Data Index: 107895 (C4 (Web Text))

Max Activating Token Index: 747

Text #13

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4307. Min Act: -0.0000

Data Index: 256238 (C4 (Web Text))

Max Activating Token Index: 328

Text #14

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4264. Min Act: -0.0000

Data Index: 1629936 (Python Code)

Max Activating Token Index: 62

Text #15

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4138. Min Act: -0.0000

Data Index: 739691 (C4 (Web Text))

Max Activating Token Index: 363

Text #16

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4137. Min Act: -0.0000

Data Index: 1365250 (Python Code)

Max Activating Token Index: 19

Text #17

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4076. Min Act: -0.0000

Data Index: 163213 (C4 (Web Text))

Max Activating Token Index: 979

Text #18

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.3893. Min Act: -0.0000

Data Index: 729752 (C4 (Web Text))

Max Activating Token Index: 1001

Text #19

Max Range: 0.5848. Min Range: -0.5848

Max Act: 0.4028. Min Act: -0.0000

Data Index: 1626773 (Python Code)

Max Activating Token Index: 1015