Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 1721 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5778. Min Act: -0.0001

Data Index: 119091 (C4 (Web Text))

Max Activating Token Index: 792

Text #1

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5587. Min Act: -0.0001

Data Index: 828109 (C4 (Web Text))

Max Activating Token Index: 157

Text #2

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5200. Min Act: -0.0001

Data Index: 176962 (C4 (Web Text))

Max Activating Token Index: 834

Text #3

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5046. Min Act: -0.0002

Data Index: 597620 (C4 (Web Text))

Max Activating Token Index: 739

Text #4

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5352. Min Act: -0.0002

Data Index: 1189024 (C4 (Web Text))

Max Activating Token Index: 830

Text #5

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5216. Min Act: -0.0001

Data Index: 492777 (C4 (Web Text))

Max Activating Token Index: 822

Text #6

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5375. Min Act: -0.0002

Data Index: 1108431 (C4 (Web Text))

Max Activating Token Index: 214

Text #7

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5316. Min Act: -0.0001

Data Index: 849962 (C4 (Web Text))

Max Activating Token Index: 471

Text #8

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5007. Min Act: -0.0001

Data Index: 540087 (C4 (Web Text))

Max Activating Token Index: 47

Text #9

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4934. Min Act: -0.0002

Data Index: 888066 (C4 (Web Text))

Max Activating Token Index: 191

Text #10

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4869. Min Act: -0.0002

Data Index: 163427 (C4 (Web Text))

Max Activating Token Index: 406

Text #11

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4984. Min Act: -0.0001

Data Index: 811231 (C4 (Web Text))

Max Activating Token Index: 253

Text #12

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4769. Min Act: -0.0001

Data Index: 1081529 (C4 (Web Text))

Max Activating Token Index: 290

Text #13

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4590. Min Act: -0.0001

Data Index: 310447 (C4 (Web Text))

Max Activating Token Index: 230

Text #14

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5120. Min Act: -0.0001

Data Index: 1538561 (Python Code)

Max Activating Token Index: 769

Text #15

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4964. Min Act: -0.0001

Data Index: 1557426 (Python Code)

Max Activating Token Index: 58

Text #16

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4564. Min Act: -0.0001

Data Index: 1377409 (Python Code)

Max Activating Token Index: 798

Text #17

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4381. Min Act: -0.0001

Data Index: 416490 (C4 (Web Text))

Max Activating Token Index: 112

Text #18

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.4694. Min Act: -0.0002

Data Index: 961520 (C4 (Web Text))

Max Activating Token Index: 103

Text #19

Max Range: 0.5778. Min Range: -0.5778

Max Act: 0.5031. Min Act: -0.0001

Data Index: 1244877 (C4 (Web Text))

Max Activating Token Index: 815