Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 765 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.8820. Min Act: -0.0000

Data Index: 1542150 (Python Code)

Max Activating Token Index: 643

Text #1

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.3380. Min Act: -0.0000

Data Index: 450616 (C4 (Web Text))

Max Activating Token Index: 993

Text #2

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.2813. Min Act: -0.0000

Data Index: 895849 (C4 (Web Text))

Max Activating Token Index: 10

Text #3

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.1183. Min Act: -0.0000

Data Index: 483358 (C4 (Web Text))

Max Activating Token Index: 560

Text #4

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.0992. Min Act: -0.0000

Data Index: 411597 (C4 (Web Text))

Max Activating Token Index: 419

Text #5

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.9215. Min Act: -0.0000

Data Index: 1075634 (C4 (Web Text))

Max Activating Token Index: 143

Text #6

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.0098. Min Act: -0.0000

Data Index: 92535 (C4 (Web Text))

Max Activating Token Index: 664

Text #7

Max Range: 3.8820. Min Range: -3.8820

Max Act: 3.0012. Min Act: -0.0000

Data Index: 37893 (C4 (Web Text))

Max Activating Token Index: 26

Text #8

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.9909. Min Act: -0.0000

Data Index: 982671 (C4 (Web Text))

Max Activating Token Index: 315

Text #9

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.9761. Min Act: -0.0000

Data Index: 390963 (C4 (Web Text))

Max Activating Token Index: 87

Text #10

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8239. Min Act: -0.0000

Data Index: 1175707 (C4 (Web Text))

Max Activating Token Index: 275

Text #11

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8973. Min Act: -0.0000

Data Index: 1444999 (Python Code)

Max Activating Token Index: 146

Text #12

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8514. Min Act: -0.0000

Data Index: 662917 (C4 (Web Text))

Max Activating Token Index: 95

Text #13

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8444. Min Act: -0.0000

Data Index: 1190279 (C4 (Web Text))

Max Activating Token Index: 135

Text #14

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8376. Min Act: -0.0000

Data Index: 420615 (C4 (Web Text))

Max Activating Token Index: 527

Text #15

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.6959. Min Act: -0.0000

Data Index: 679686 (C4 (Web Text))

Max Activating Token Index: 576

Text #16

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8378. Min Act: -0.0000

Data Index: 1565672 (Python Code)

Max Activating Token Index: 415

Text #17

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.8025. Min Act: -0.0000

Data Index: 1544798 (Python Code)

Max Activating Token Index: 82

Text #18

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.6537. Min Act: -0.0000

Data Index: 1397660 (Python Code)

Max Activating Token Index: 180

Text #19

Max Range: 3.8820. Min Range: -3.8820

Max Act: 2.7898. Min Act: -0.0000

Data Index: 407701 (C4 (Web Text))

Max Activating Token Index: 462