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

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

Neuron 4632 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.7897. Min Act: -0.0000

Data Index: 354746 (C4 (Web Text))

Max Activating Token Index: 781

Text #1

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.6393. Min Act: -0.0000

Data Index: 1618823 (Python Code)

Max Activating Token Index: 163

Text #2

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5869. Min Act: -0.0000

Data Index: 936425 (C4 (Web Text))

Max Activating Token Index: 76

Text #3

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.6072. Min Act: -0.0000

Data Index: 741103 (C4 (Web Text))

Max Activating Token Index: 309

Text #4

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.6008. Min Act: -0.0000

Data Index: 199061 (C4 (Web Text))

Max Activating Token Index: 267

Text #5

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5666. Min Act: -0.0000

Data Index: 168874 (C4 (Web Text))

Max Activating Token Index: 1015

Text #6

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5645. Min Act: -0.0000

Data Index: 1192915 (C4 (Web Text))

Max Activating Token Index: 270

Text #7

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5353. Min Act: -0.0000

Data Index: 969785 (C4 (Web Text))

Max Activating Token Index: 421

Text #8

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5254. Min Act: -0.0000

Data Index: 1180512 (C4 (Web Text))

Max Activating Token Index: 252

Text #9

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5240. Min Act: -0.0000

Data Index: 1345575 (C4 (Web Text))

Max Activating Token Index: 505

Text #10

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5292. Min Act: -0.0000

Data Index: 916017 (C4 (Web Text))

Max Activating Token Index: 383

Text #11

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4934. Min Act: -0.0000

Data Index: 1643855 (Python Code)

Max Activating Token Index: 50

Text #12

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.5026. Min Act: -0.0000

Data Index: 1138809 (C4 (Web Text))

Max Activating Token Index: 1014

Text #13

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4586. Min Act: -0.0000

Data Index: 1202488 (C4 (Web Text))

Max Activating Token Index: 717

Text #14

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4221. Min Act: -0.0000

Data Index: 1467678 (Python Code)

Max Activating Token Index: 439

Text #15

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4308. Min Act: -0.0000

Data Index: 229631 (C4 (Web Text))

Max Activating Token Index: 397

Text #16

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4271. Min Act: -0.0000

Data Index: 284696 (C4 (Web Text))

Max Activating Token Index: 76

Text #17

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4300. Min Act: -0.0000

Data Index: 108621 (C4 (Web Text))

Max Activating Token Index: 388

Text #18

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4496. Min Act: -0.0000

Data Index: 338457 (C4 (Web Text))

Max Activating Token Index: 374

Text #19

Max Range: 0.7897. Min Range: -0.7897

Max Act: 0.4468. Min Act: -0.0000

Data Index: 40426 (C4 (Web Text))

Max Activating Token Index: 752