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

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

Neuron 3366 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2740. Min Act: -0.0001

Data Index: 125826 (C4 (Web Text))

Max Activating Token Index: 698

Text #1

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2770. Min Act: -0.0000

Data Index: 1065139 (C4 (Web Text))

Max Activating Token Index: 303

Text #2

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2811. Min Act: -0.0000

Data Index: 1317939 (C4 (Web Text))

Max Activating Token Index: 120

Text #3

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2710. Min Act: -0.0000

Data Index: 720650 (C4 (Web Text))

Max Activating Token Index: 614

Text #4

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2803. Min Act: -0.0000

Data Index: 1165098 (C4 (Web Text))

Max Activating Token Index: 141

Text #5

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2560. Min Act: -0.0001

Data Index: 768779 (C4 (Web Text))

Max Activating Token Index: 274

Text #6

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2599. Min Act: -0.0000

Data Index: 777009 (C4 (Web Text))

Max Activating Token Index: 60

Text #7

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2441. Min Act: -0.0001

Data Index: 1062492 (C4 (Web Text))

Max Activating Token Index: 604

Text #8

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2318. Min Act: -0.0000

Data Index: 813748 (C4 (Web Text))

Max Activating Token Index: 575

Text #9

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2379. Min Act: -0.0000

Data Index: 114310 (C4 (Web Text))

Max Activating Token Index: 700

Text #10

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2381. Min Act: -0.0000

Data Index: 935089 (C4 (Web Text))

Max Activating Token Index: 466

Text #11

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2372. Min Act: -0.0001

Data Index: 887457 (C4 (Web Text))

Max Activating Token Index: 96

Text #12

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2287. Min Act: -0.0000

Data Index: 1174857 (C4 (Web Text))

Max Activating Token Index: 824

Text #13

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2407. Min Act: -0.0000

Data Index: 1578699 (Python Code)

Max Activating Token Index: 330

Text #14

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2174. Min Act: -0.0000

Data Index: 882898 (C4 (Web Text))

Max Activating Token Index: 817

Text #15

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2209. Min Act: -0.0001

Data Index: 964051 (C4 (Web Text))

Max Activating Token Index: 819

Text #16

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2105. Min Act: -0.0001

Data Index: 1249193 (C4 (Web Text))

Max Activating Token Index: 885

Text #17

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2193. Min Act: -0.0000

Data Index: 695543 (C4 (Web Text))

Max Activating Token Index: 128

Text #18

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2192. Min Act: -0.0000

Data Index: 774725 (C4 (Web Text))

Max Activating Token Index: 801

Text #19

Max Range: 0.2811. Min Range: -0.2811

Max Act: 0.2105. Min Act: -0.0001

Data Index: 2712 (C4 (Web Text))

Max Activating Token Index: 566