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

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

Neuron 2475 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.5764. Min Act: -0.0001

Data Index: 1219323 (C4 (Web Text))

Max Activating Token Index: 96

Text #1

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.4523. Min Act: -0.0001

Data Index: 173571 (C4 (Web Text))

Max Activating Token Index: 140

Text #2

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.4517. Min Act: -0.0001

Data Index: 1344749 (C4 (Web Text))

Max Activating Token Index: 73

Text #3

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.4330. Min Act: -0.0001

Data Index: 838993 (C4 (Web Text))

Max Activating Token Index: 1020

Text #4

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.4088. Min Act: -0.0001

Data Index: 655050 (C4 (Web Text))

Max Activating Token Index: 541

Text #5

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3920. Min Act: -0.0001

Data Index: 483539 (C4 (Web Text))

Max Activating Token Index: 841

Text #6

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.4186. Min Act: -0.0001

Data Index: 1350646 (C4 (Web Text))

Max Activating Token Index: 252

Text #7

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3919. Min Act: -0.0001

Data Index: 190373 (C4 (Web Text))

Max Activating Token Index: 154

Text #8

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3802. Min Act: -0.0001

Data Index: 704119 (C4 (Web Text))

Max Activating Token Index: 929

Text #9

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3886. Min Act: -0.0001

Data Index: 115984 (C4 (Web Text))

Max Activating Token Index: 612

Text #10

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3812. Min Act: -0.0001

Data Index: 1305065 (C4 (Web Text))

Max Activating Token Index: 138

Text #11

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3872. Min Act: -0.0001

Data Index: 936344 (C4 (Web Text))

Max Activating Token Index: 921

Text #12

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3619. Min Act: -0.0001

Data Index: 308767 (C4 (Web Text))

Max Activating Token Index: 679

Text #13

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3756. Min Act: -0.0001

Data Index: 633167 (C4 (Web Text))

Max Activating Token Index: 931

Text #14

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3667. Min Act: -0.0001

Data Index: 1256910 (C4 (Web Text))

Max Activating Token Index: 906

Text #15

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3641. Min Act: -0.0001

Data Index: 471050 (C4 (Web Text))

Max Activating Token Index: 880

Text #16

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3495. Min Act: -0.0001

Data Index: 1573548 (Python Code)

Max Activating Token Index: 351

Text #17

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3404. Min Act: -0.0001

Data Index: 1151263 (C4 (Web Text))

Max Activating Token Index: 181

Text #18

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3600. Min Act: -0.0001

Data Index: 756823 (C4 (Web Text))

Max Activating Token Index: 605

Text #19

Max Range: 0.5764. Min Range: -0.5764

Max Act: 0.3552. Min Act: -0.0001

Data Index: 500098 (C4 (Web Text))

Max Activating Token Index: 283