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

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

Neuron 2900 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.2898. Min Range: -5.2898

Max Act: 5.2898. Min Act: -0.0000

Data Index: 208798 (C4 (Web Text))

Max Activating Token Index: 268

Text #1

Max Range: 5.2898. Min Range: -5.2898

Max Act: 4.8771. Min Act: -0.0000

Data Index: 46119 (C4 (Web Text))

Max Activating Token Index: 920

Text #2

Max Range: 5.2898. Min Range: -5.2898

Max Act: 4.1144. Min Act: -0.0000

Data Index: 1035572 (C4 (Web Text))

Max Activating Token Index: 797

Text #3

Max Range: 5.2898. Min Range: -5.2898

Max Act: 4.0634. Min Act: -0.0000

Data Index: 293830 (C4 (Web Text))

Max Activating Token Index: 160

Text #4

Max Range: 5.2898. Min Range: -5.2898

Max Act: 4.1909. Min Act: -0.0000

Data Index: 74342 (C4 (Web Text))

Max Activating Token Index: 762

Text #5

Max Range: 5.2898. Min Range: -5.2898

Max Act: 4.4296. Min Act: -0.0000

Data Index: 907025 (C4 (Web Text))

Max Activating Token Index: 652

Text #6

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.4887. Min Act: -0.0000

Data Index: 586415 (C4 (Web Text))

Max Activating Token Index: 79

Text #7

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.8141. Min Act: -0.0000

Data Index: 676778 (C4 (Web Text))

Max Activating Token Index: 66

Text #8

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.6810. Min Act: -0.0000

Data Index: 1078003 (C4 (Web Text))

Max Activating Token Index: 339

Text #9

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.7464. Min Act: -0.0000

Data Index: 741726 (C4 (Web Text))

Max Activating Token Index: 585

Text #10

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.5874. Min Act: -0.0000

Data Index: 1116464 (C4 (Web Text))

Max Activating Token Index: 285

Text #11

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.9919. Min Act: -0.0000

Data Index: 213490 (C4 (Web Text))

Max Activating Token Index: 521

Text #12

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.9740. Min Act: -0.0000

Data Index: 1207607 (C4 (Web Text))

Max Activating Token Index: 72

Text #13

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.7728. Min Act: -0.0000

Data Index: 1642373 (Python Code)

Max Activating Token Index: 462

Text #14

Max Range: 5.2898. Min Range: -5.2898

Max Act: 3.0388. Min Act: -0.0000

Data Index: 1502657 (Python Code)

Max Activating Token Index: 567

Text #15

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.8569. Min Act: -0.0000

Data Index: 1614080 (Python Code)

Max Activating Token Index: 180

Text #16

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.5047. Min Act: -0.0000

Data Index: 234680 (C4 (Web Text))

Max Activating Token Index: 61

Text #17

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.6036. Min Act: -0.0000

Data Index: 1503372 (Python Code)

Max Activating Token Index: 333

Text #18

Max Range: 5.2898. Min Range: -5.2898

Max Act: 2.0980. Min Act: -0.0000

Data Index: 1493087 (Python Code)

Max Activating Token Index: 628

Text #19

Max Range: 5.2898. Min Range: -5.2898

Max Act: 1.9726. Min Act: -0.0000

Data Index: 1480181 (Python Code)

Max Activating Token Index: 502