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

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

Neuron 595 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.6444. Min Act: -0.0001

Data Index: 1564153 (Python Code)

Max Activating Token Index: 442

Text #1

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.6852. Min Act: -0.0000

Data Index: 1112117 (C4 (Web Text))

Max Activating Token Index: 219

Text #2

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5790. Min Act: -0.0001

Data Index: 483173 (C4 (Web Text))

Max Activating Token Index: 906

Text #3

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.6107. Min Act: -0.0001

Data Index: 1634612 (Python Code)

Max Activating Token Index: 364

Text #4

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5740. Min Act: -0.0000

Data Index: 261835 (C4 (Web Text))

Max Activating Token Index: 924

Text #5

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5678. Min Act: -0.0001

Data Index: 1534795 (Python Code)

Max Activating Token Index: 899

Text #6

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5547. Min Act: -0.0001

Data Index: 43914 (C4 (Web Text))

Max Activating Token Index: 183

Text #7

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.6045. Min Act: -0.0001

Data Index: 488645 (C4 (Web Text))

Max Activating Token Index: 258

Text #8

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5660. Min Act: -0.0001

Data Index: 805922 (C4 (Web Text))

Max Activating Token Index: 391

Text #9

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5430. Min Act: -0.0000

Data Index: 725697 (C4 (Web Text))

Max Activating Token Index: 402

Text #10

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5418. Min Act: -0.0000

Data Index: 120584 (C4 (Web Text))

Max Activating Token Index: 794

Text #11

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5788. Min Act: -0.0001

Data Index: 1588609 (Python Code)

Max Activating Token Index: 921

Text #12

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5369. Min Act: -0.0001

Data Index: 577933 (C4 (Web Text))

Max Activating Token Index: 951

Text #13

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5422. Min Act: -0.0001

Data Index: 671374 (C4 (Web Text))

Max Activating Token Index: 748

Text #14

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5512. Min Act: -0.0000

Data Index: 511550 (C4 (Web Text))

Max Activating Token Index: 116

Text #15

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5449. Min Act: -0.0001

Data Index: 1507657 (Python Code)

Max Activating Token Index: 686

Text #16

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5176. Min Act: -0.0000

Data Index: 489038 (C4 (Web Text))

Max Activating Token Index: 669

Text #17

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5289. Min Act: -0.0001

Data Index: 250163 (C4 (Web Text))

Max Activating Token Index: 467

Text #18

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5278. Min Act: -0.0000

Data Index: 675275 (C4 (Web Text))

Max Activating Token Index: 596

Text #19

Max Range: 0.6852. Min Range: -0.6852

Max Act: 0.5074. Min Act: -0.0001

Data Index: 806479 (C4 (Web Text))

Max Activating Token Index: 466