Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 216 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0740. Min Act: -0.0003

Data Index: 373808 (C4 (Web Text))

Max Activating Token Index: 469

Text #1

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0707. Min Act: -0.0003

Data Index: 598689 (C4 (Web Text))

Max Activating Token Index: 23

Text #2

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0707. Min Act: -0.0003

Data Index: 1319933 (C4 (Web Text))

Max Activating Token Index: 483

Text #3

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0690. Min Act: -0.0004

Data Index: 275787 (C4 (Web Text))

Max Activating Token Index: 157

Text #4

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0696. Min Act: -0.0003

Data Index: 127131 (C4 (Web Text))

Max Activating Token Index: 552

Text #5

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0682. Min Act: -0.0003

Data Index: 539132 (C4 (Web Text))

Max Activating Token Index: 230

Text #6

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0679. Min Act: -0.0004

Data Index: 995963 (C4 (Web Text))

Max Activating Token Index: 557

Text #7

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0696. Min Act: -0.0003

Data Index: 458363 (C4 (Web Text))

Max Activating Token Index: 988

Text #8

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0663. Min Act: -0.0003

Data Index: 1008427 (C4 (Web Text))

Max Activating Token Index: 189

Text #9

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0662. Min Act: -0.0003

Data Index: 474185 (C4 (Web Text))

Max Activating Token Index: 154

Text #10

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0651. Min Act: -0.0003

Data Index: 1235424 (C4 (Web Text))

Max Activating Token Index: 812

Text #11

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0680. Min Act: -0.0003

Data Index: 1233353 (C4 (Web Text))

Max Activating Token Index: 792

Text #12

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0666. Min Act: -0.0003

Data Index: 134644 (C4 (Web Text))

Max Activating Token Index: 462

Text #13

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0663. Min Act: -0.0004

Data Index: 9700 (C4 (Web Text))

Max Activating Token Index: 381

Text #14

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0654. Min Act: -0.0003

Data Index: 474673 (C4 (Web Text))

Max Activating Token Index: 555

Text #15

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0670. Min Act: -0.0003

Data Index: 294301 (C4 (Web Text))

Max Activating Token Index: 128

Text #16

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0668. Min Act: -0.0002

Data Index: 532787 (C4 (Web Text))

Max Activating Token Index: 500

Text #17

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0668. Min Act: -0.0003

Data Index: 387651 (C4 (Web Text))

Max Activating Token Index: 1003

Text #18

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0667. Min Act: -0.0003

Data Index: 1469683 (Python Code)

Max Activating Token Index: 122

Text #19

Max Range: 0.0740. Min Range: -0.0740

Max Act: 0.0627. Min Act: -0.0003

Data Index: 529774 (C4 (Web Text))

Max Activating Token Index: 978