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

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

Neuron 3962 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: 0.0714. Min Range: -0.0714

Max Act: 0.0711. Min Act: -0.0001

Data Index: 1054770 (C4 (Web Text))

Max Activating Token Index: 711

Text #1

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0714. Min Act: -0.0001

Data Index: 208207 (C4 (Web Text))

Max Activating Token Index: 229

Text #2

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0701. Min Act: -0.0001

Data Index: 634207 (C4 (Web Text))

Max Activating Token Index: 322

Text #3

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0634. Min Act: -0.0001

Data Index: 582735 (C4 (Web Text))

Max Activating Token Index: 147

Text #4

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0636. Min Act: -0.0001

Data Index: 884622 (C4 (Web Text))

Max Activating Token Index: 93

Text #5

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0601. Min Act: -0.0001

Data Index: 997535 (C4 (Web Text))

Max Activating Token Index: 164

Text #6

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0570. Min Act: -0.0001

Data Index: 421459 (C4 (Web Text))

Max Activating Token Index: 984

Text #7

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0599. Min Act: -0.0001

Data Index: 970952 (C4 (Web Text))

Max Activating Token Index: 124

Text #8

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0593. Min Act: -0.0001

Data Index: 652364 (C4 (Web Text))

Max Activating Token Index: 880

Text #9

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0547. Min Act: -0.0001

Data Index: 1431810 (Python Code)

Max Activating Token Index: 850

Text #10

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0540. Min Act: -0.0001

Data Index: 367484 (C4 (Web Text))

Max Activating Token Index: 715

Text #11

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0507. Min Act: -0.0001

Data Index: 1197720 (C4 (Web Text))

Max Activating Token Index: 261

Text #12

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0541. Min Act: -0.0001

Data Index: 887837 (C4 (Web Text))

Max Activating Token Index: 261

Text #13

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0501. Min Act: -0.0001

Data Index: 82369 (C4 (Web Text))

Max Activating Token Index: 692

Text #14

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0496. Min Act: -0.0001

Data Index: 651386 (C4 (Web Text))

Max Activating Token Index: 512

Text #15

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0491. Min Act: -0.0001

Data Index: 1003062 (C4 (Web Text))

Max Activating Token Index: 598

Text #16

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0483. Min Act: -0.0001

Data Index: 1524862 (Python Code)

Max Activating Token Index: 566

Text #17

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0503. Min Act: -0.0001

Data Index: 550788 (C4 (Web Text))

Max Activating Token Index: 246

Text #18

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0490. Min Act: -0.0001

Data Index: 124510 (C4 (Web Text))

Max Activating Token Index: 447

Text #19

Max Range: 0.0714. Min Range: -0.0714

Max Act: 0.0470. Min Act: -0.0001

Data Index: 1648460 (Python Code)

Max Activating Token Index: 904