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

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

Neuron 2512 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4749. Min Act: -0.0000

Data Index: 642105 (C4 (Web Text))

Max Activating Token Index: 549

Text #1

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4276. Min Act: -0.0000

Data Index: 742664 (C4 (Web Text))

Max Activating Token Index: 229

Text #2

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4439. Min Act: -0.0000

Data Index: 195280 (C4 (Web Text))

Max Activating Token Index: 850

Text #3

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4154. Min Act: -0.0000

Data Index: 120975 (C4 (Web Text))

Max Activating Token Index: 530

Text #4

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4122. Min Act: -0.0000

Data Index: 368407 (C4 (Web Text))

Max Activating Token Index: 268

Text #5

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4176. Min Act: -0.0000

Data Index: 51623 (C4 (Web Text))

Max Activating Token Index: 858

Text #6

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4106. Min Act: -0.0000

Data Index: 504822 (C4 (Web Text))

Max Activating Token Index: 84

Text #7

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4090. Min Act: -0.0000

Data Index: 75850 (C4 (Web Text))

Max Activating Token Index: 321

Text #8

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.4179. Min Act: -0.0000

Data Index: 596857 (C4 (Web Text))

Max Activating Token Index: 772

Text #9

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3686. Min Act: -0.0000

Data Index: 643201 (C4 (Web Text))

Max Activating Token Index: 423

Text #10

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3480. Min Act: -0.0000

Data Index: 91394 (C4 (Web Text))

Max Activating Token Index: 274

Text #11

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3479. Min Act: -0.0000

Data Index: 959240 (C4 (Web Text))

Max Activating Token Index: 125

Text #12

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3250. Min Act: -0.0000

Data Index: 748232 (C4 (Web Text))

Max Activating Token Index: 757

Text #13

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3779. Min Act: -0.0000

Data Index: 589731 (C4 (Web Text))

Max Activating Token Index: 1020

Text #14

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3229. Min Act: -0.0000

Data Index: 1516337 (Python Code)

Max Activating Token Index: 782

Text #15

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3077. Min Act: -0.0000

Data Index: 1640361 (Python Code)

Max Activating Token Index: 519

Text #16

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3098. Min Act: -0.0000

Data Index: 958164 (C4 (Web Text))

Max Activating Token Index: 271

Text #17

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3161. Min Act: -0.0000

Data Index: 586956 (C4 (Web Text))

Max Activating Token Index: 790

Text #18

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3091. Min Act: -0.0000

Data Index: 1473108 (Python Code)

Max Activating Token Index: 665

Text #19

Max Range: 0.4749. Min Range: -0.4749

Max Act: 0.3098. Min Act: -0.0000

Data Index: 1021811 (C4 (Web Text))

Max Activating Token Index: 618