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

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

Neuron 4227 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: 3.3021. Min Range: -3.3021

Max Act: 3.3021. Min Act: -0.0001

Data Index: 323457 (C4 (Web Text))

Max Activating Token Index: 46

Text #1

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.6950. Min Act: -0.0001

Data Index: 1285418 (C4 (Web Text))

Max Activating Token Index: 864

Text #2

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.9884. Min Act: -0.0001

Data Index: 869501 (C4 (Web Text))

Max Activating Token Index: 764

Text #3

Max Range: 3.3021. Min Range: -3.3021

Max Act: 3.0435. Min Act: -0.0001

Data Index: 1503422 (Python Code)

Max Activating Token Index: 637

Text #4

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.8262. Min Act: -0.0001

Data Index: 1640179 (Python Code)

Max Activating Token Index: 526

Text #5

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.9684. Min Act: -0.0001

Data Index: 1473648 (Python Code)

Max Activating Token Index: 141

Text #6

Max Range: 3.3021. Min Range: -3.3021

Max Act: 3.1218. Min Act: -0.0001

Data Index: 1135719 (C4 (Web Text))

Max Activating Token Index: 568

Text #7

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.9192. Min Act: -0.0001

Data Index: 1597312 (Python Code)

Max Activating Token Index: 797

Text #8

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.9241. Min Act: -0.0001

Data Index: 209904 (C4 (Web Text))

Max Activating Token Index: 56

Text #9

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.7964. Min Act: -0.0001

Data Index: 579204 (C4 (Web Text))

Max Activating Token Index: 674

Text #10

Max Range: 3.3021. Min Range: -3.3021

Max Act: 3.0384. Min Act: -0.0001

Data Index: 1262314 (C4 (Web Text))

Max Activating Token Index: 914

Text #11

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.9014. Min Act: -0.0001

Data Index: 1451232 (Python Code)

Max Activating Token Index: 52

Text #12

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.8151. Min Act: -0.0001

Data Index: 1518822 (Python Code)

Max Activating Token Index: 37

Text #13

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.8163. Min Act: -0.0001

Data Index: 702223 (C4 (Web Text))

Max Activating Token Index: 832

Text #14

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.5435. Min Act: -0.0001

Data Index: 822747 (C4 (Web Text))

Max Activating Token Index: 266

Text #15

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.8717. Min Act: -0.0001

Data Index: 1128053 (C4 (Web Text))

Max Activating Token Index: 89

Text #16

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.7519. Min Act: -0.0001

Data Index: 1608154 (Python Code)

Max Activating Token Index: 918

Text #17

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.7824. Min Act: -0.0001

Data Index: 1561034 (Python Code)

Max Activating Token Index: 61

Text #18

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.7495. Min Act: -0.0001

Data Index: 416921 (C4 (Web Text))

Max Activating Token Index: 683

Text #19

Max Range: 3.3021. Min Range: -3.3021

Max Act: 2.6005. Min Act: -0.0001

Data Index: 1539488 (Python Code)

Max Activating Token Index: 54