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

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

Neuron 1040 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: 0.4666. Min Range: -0.4666

Max Act: 0.4666. Min Act: -0.0000

Data Index: 1286540 (C4 (Web Text))

Max Activating Token Index: 734

Text #1

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3797. Min Act: -0.0000

Data Index: 882164 (C4 (Web Text))

Max Activating Token Index: 188

Text #2

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3507. Min Act: -0.0000

Data Index: 319073 (C4 (Web Text))

Max Activating Token Index: 461

Text #3

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3028. Min Act: -0.0000

Data Index: 231269 (C4 (Web Text))

Max Activating Token Index: 156

Text #4

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3146. Min Act: -0.0000

Data Index: 667958 (C4 (Web Text))

Max Activating Token Index: 386

Text #5

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3331. Min Act: -0.0000

Data Index: 1069025 (C4 (Web Text))

Max Activating Token Index: 226

Text #6

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3309. Min Act: -0.0000

Data Index: 175585 (C4 (Web Text))

Max Activating Token Index: 396

Text #7

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3129. Min Act: -0.0000

Data Index: 1551332 (Python Code)

Max Activating Token Index: 141

Text #8

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2925. Min Act: -0.0000

Data Index: 1629160 (Python Code)

Max Activating Token Index: 929

Text #9

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3053. Min Act: -0.0000

Data Index: 1270659 (C4 (Web Text))

Max Activating Token Index: 189

Text #10

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.3103. Min Act: -0.0000

Data Index: 672118 (C4 (Web Text))

Max Activating Token Index: 442

Text #11

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2862. Min Act: -0.0000

Data Index: 1469162 (Python Code)

Max Activating Token Index: 44

Text #12

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2940. Min Act: -0.0000

Data Index: 1077257 (C4 (Web Text))

Max Activating Token Index: 706

Text #13

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2936. Min Act: -0.0000

Data Index: 90329 (C4 (Web Text))

Max Activating Token Index: 344

Text #14

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2704. Min Act: -0.0000

Data Index: 1511583 (Python Code)

Max Activating Token Index: 172

Text #15

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2813. Min Act: -0.0000

Data Index: 329089 (C4 (Web Text))

Max Activating Token Index: 830

Text #16

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2749. Min Act: -0.0000

Data Index: 1605952 (Python Code)

Max Activating Token Index: 823

Text #17

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2423. Min Act: -0.0000

Data Index: 1445941 (Python Code)

Max Activating Token Index: 93

Text #18

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2642. Min Act: -0.0000

Data Index: 785273 (C4 (Web Text))

Max Activating Token Index: 375

Text #19

Max Range: 0.4666. Min Range: -0.4666

Max Act: 0.2340. Min Act: -0.0000

Data Index: 863663 (C4 (Web Text))

Max Activating Token Index: 330