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

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

Neuron 1617 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.9280. Min Act: -0.0001

Data Index: 394526 (C4 (Web Text))

Max Activating Token Index: 217

Text #1

Max Range: 4.1059. Min Range: -4.1059

Max Act: 4.1059. Min Act: -0.0001

Data Index: 1654198 (Python Code)

Max Activating Token Index: 146

Text #2

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.8686. Min Act: -0.0001

Data Index: 640766 (C4 (Web Text))

Max Activating Token Index: 119

Text #3

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.7304. Min Act: -0.0001

Data Index: 1416374 (Python Code)

Max Activating Token Index: 691

Text #4

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.9543. Min Act: -0.0001

Data Index: 1295171 (C4 (Web Text))

Max Activating Token Index: 130

Text #5

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.7489. Min Act: -0.0000

Data Index: 80674 (C4 (Web Text))

Max Activating Token Index: 869

Text #6

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.4242. Min Act: -0.0001

Data Index: 1377064 (Python Code)

Max Activating Token Index: 17

Text #7

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.3861. Min Act: -0.0001

Data Index: 239455 (C4 (Web Text))

Max Activating Token Index: 193

Text #8

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.4975. Min Act: -0.0000

Data Index: 855475 (C4 (Web Text))

Max Activating Token Index: 324

Text #9

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.1743. Min Act: -0.0001

Data Index: 157635 (C4 (Web Text))

Max Activating Token Index: 253

Text #10

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.2797. Min Act: -0.0001

Data Index: 331044 (C4 (Web Text))

Max Activating Token Index: 984

Text #11

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.2873. Min Act: -0.0001

Data Index: 1386078 (Python Code)

Max Activating Token Index: 35

Text #12

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.4257. Min Act: -0.0001

Data Index: 1520038 (Python Code)

Max Activating Token Index: 52

Text #13

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.0911. Min Act: -0.0001

Data Index: 1193044 (C4 (Web Text))

Max Activating Token Index: 640

Text #14

Max Range: 4.1059. Min Range: -4.1059

Max Act: 2.9045. Min Act: -0.0001

Data Index: 1477257 (Python Code)

Max Activating Token Index: 73

Text #15

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.3386. Min Act: -0.0001

Data Index: 1492367 (Python Code)

Max Activating Token Index: 120

Text #16

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.6027. Min Act: -0.0001

Data Index: 347711 (C4 (Web Text))

Max Activating Token Index: 494

Text #17

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.5651. Min Act: -0.0001

Data Index: 776144 (C4 (Web Text))

Max Activating Token Index: 242

Text #18

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.2765. Min Act: -0.0000

Data Index: 1612534 (Python Code)

Max Activating Token Index: 34

Text #19

Max Range: 4.1059. Min Range: -4.1059

Max Act: 3.5686. Min Act: -0.0001

Data Index: 766392 (C4 (Web Text))

Max Activating Token Index: 524