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

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

Neuron 2795 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2259. Min Act: -0.0001

Data Index: 133484 (C4 (Web Text))

Max Activating Token Index: 773

Text #1

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2339. Min Act: -0.0001

Data Index: 719964 (C4 (Web Text))

Max Activating Token Index: 1005

Text #2

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2186. Min Act: -0.0001

Data Index: 629238 (C4 (Web Text))

Max Activating Token Index: 281

Text #3

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2244. Min Act: -0.0001

Data Index: 750966 (C4 (Web Text))

Max Activating Token Index: 399

Text #4

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2165. Min Act: -0.0001

Data Index: 637138 (C4 (Web Text))

Max Activating Token Index: 890

Text #5

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2219. Min Act: -0.0001

Data Index: 144642 (C4 (Web Text))

Max Activating Token Index: 76

Text #6

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2111. Min Act: -0.0001

Data Index: 1302274 (C4 (Web Text))

Max Activating Token Index: 622

Text #7

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2096. Min Act: -0.0001

Data Index: 133375 (C4 (Web Text))

Max Activating Token Index: 379

Text #8

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2026. Min Act: -0.0001

Data Index: 74539 (C4 (Web Text))

Max Activating Token Index: 329

Text #9

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1990. Min Act: -0.0001

Data Index: 334498 (C4 (Web Text))

Max Activating Token Index: 849

Text #10

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1972. Min Act: -0.0001

Data Index: 266799 (C4 (Web Text))

Max Activating Token Index: 693

Text #11

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2010. Min Act: -0.0001

Data Index: 39330 (C4 (Web Text))

Max Activating Token Index: 553

Text #12

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1906. Min Act: -0.0001

Data Index: 512755 (C4 (Web Text))

Max Activating Token Index: 303

Text #13

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2142. Min Act: -0.0001

Data Index: 290219 (C4 (Web Text))

Max Activating Token Index: 828

Text #14

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1879. Min Act: -0.0001

Data Index: 710327 (C4 (Web Text))

Max Activating Token Index: 1013

Text #15

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2121. Min Act: -0.0001

Data Index: 1478415 (Python Code)

Max Activating Token Index: 768

Text #16

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1856. Min Act: -0.0001

Data Index: 1097943 (C4 (Web Text))

Max Activating Token Index: 41

Text #17

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.2149. Min Act: -0.0001

Data Index: 45038 (C4 (Web Text))

Max Activating Token Index: 337

Text #18

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1945. Min Act: -0.0001

Data Index: 1312588 (C4 (Web Text))

Max Activating Token Index: 358

Text #19

Max Range: 0.2339. Min Range: -0.2339

Max Act: 0.1901. Min Act: -0.0001

Data Index: 803925 (C4 (Web Text))

Max Activating Token Index: 692