Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 768 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 2.3137. Min Range: -2.3137

Max Act: 2.3137. Min Act: -0.0000

Data Index: 1027358 (C4 (Web Text))

Max Activating Token Index: 993

Text #1

Max Range: 2.3137. Min Range: -2.3137

Max Act: 2.2435. Min Act: -0.0000

Data Index: 641223 (C4 (Web Text))

Max Activating Token Index: 630

Text #2

Max Range: 2.3137. Min Range: -2.3137

Max Act: 2.1162. Min Act: -0.0000

Data Index: 445582 (C4 (Web Text))

Max Activating Token Index: 153

Text #3

Max Range: 2.3137. Min Range: -2.3137

Max Act: 2.1783. Min Act: -0.0000

Data Index: 207790 (C4 (Web Text))

Max Activating Token Index: 493

Text #4

Max Range: 2.3137. Min Range: -2.3137

Max Act: 2.0413. Min Act: -0.0000

Data Index: 518621 (C4 (Web Text))

Max Activating Token Index: 473

Text #5

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.9228. Min Act: -0.0000

Data Index: 541723 (C4 (Web Text))

Max Activating Token Index: 968

Text #6

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.8211. Min Act: -0.0001

Data Index: 826125 (C4 (Web Text))

Max Activating Token Index: 788

Text #7

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.9472. Min Act: -0.0000

Data Index: 1342553 (C4 (Web Text))

Max Activating Token Index: 541

Text #8

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.8535. Min Act: -0.0000

Data Index: 474310 (C4 (Web Text))

Max Activating Token Index: 648

Text #9

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.9798. Min Act: -0.0001

Data Index: 1016611 (C4 (Web Text))

Max Activating Token Index: 804

Text #10

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.8407. Min Act: -0.0000

Data Index: 12638 (C4 (Web Text))

Max Activating Token Index: 488

Text #11

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.9749. Min Act: -0.0001

Data Index: 1419952 (Python Code)

Max Activating Token Index: 626

Text #12

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.5099. Min Act: -0.0001

Data Index: 1459500 (Python Code)

Max Activating Token Index: 163

Text #13

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.8392. Min Act: -0.0000

Data Index: 1361849 (Python Code)

Max Activating Token Index: 1002

Text #14

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.7040. Min Act: -0.0000

Data Index: 1056995 (C4 (Web Text))

Max Activating Token Index: 195

Text #15

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.7516. Min Act: -0.0000

Data Index: 1567364 (Python Code)

Max Activating Token Index: 891

Text #16

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.7414. Min Act: -0.0001

Data Index: 1452232 (Python Code)

Max Activating Token Index: 812

Text #17

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.6199. Min Act: -0.0001

Data Index: 1380957 (Python Code)

Max Activating Token Index: 1017

Text #18

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.7875. Min Act: -0.0000

Data Index: 1279659 (C4 (Web Text))

Max Activating Token Index: 481

Text #19

Max Range: 2.3137. Min Range: -2.3137

Max Act: 1.8673. Min Act: -0.0000

Data Index: 488965 (C4 (Web Text))

Max Activating Token Index: 201