Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 2704 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.9146. Min Act: -0.0001

Data Index: 172685 (C4 (Web Text))

Max Activating Token Index: 21

Text #1

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7637. Min Act: -0.0001

Data Index: 25549 (C4 (Web Text))

Max Activating Token Index: 515

Text #2

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7394. Min Act: -0.0001

Data Index: 336500 (C4 (Web Text))

Max Activating Token Index: 274

Text #3

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7213. Min Act: -0.0001

Data Index: 171876 (C4 (Web Text))

Max Activating Token Index: 57

Text #4

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6791. Min Act: -0.0001

Data Index: 483380 (C4 (Web Text))

Max Activating Token Index: 743

Text #5

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7378. Min Act: -0.0001

Data Index: 592807 (C4 (Web Text))

Max Activating Token Index: 886

Text #6

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7343. Min Act: -0.0001

Data Index: 700049 (C4 (Web Text))

Max Activating Token Index: 55

Text #7

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7279. Min Act: -0.0001

Data Index: 626410 (C4 (Web Text))

Max Activating Token Index: 1009

Text #8

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7391. Min Act: -0.0001

Data Index: 1520542 (Python Code)

Max Activating Token Index: 211

Text #9

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6854. Min Act: -0.0001

Data Index: 607119 (C4 (Web Text))

Max Activating Token Index: 321

Text #10

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7058. Min Act: -0.0001

Data Index: 861269 (C4 (Web Text))

Max Activating Token Index: 879

Text #11

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7272. Min Act: -0.0001

Data Index: 777064 (C4 (Web Text))

Max Activating Token Index: 192

Text #12

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6823. Min Act: -0.0001

Data Index: 63637 (C4 (Web Text))

Max Activating Token Index: 903

Text #13

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6878. Min Act: -0.0001

Data Index: 165520 (C4 (Web Text))

Max Activating Token Index: 798

Text #14

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.7099. Min Act: -0.0001

Data Index: 1192223 (C4 (Web Text))

Max Activating Token Index: 986

Text #15

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6896. Min Act: -0.0001

Data Index: 146055 (C4 (Web Text))

Max Activating Token Index: 7

Text #16

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6451. Min Act: -0.0001

Data Index: 1505766 (Python Code)

Max Activating Token Index: 768

Text #17

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6606. Min Act: -0.0001

Data Index: 320113 (C4 (Web Text))

Max Activating Token Index: 1011

Text #18

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6779. Min Act: -0.0001

Data Index: 1539346 (Python Code)

Max Activating Token Index: 259

Text #19

Max Range: 0.9146. Min Range: -0.9146

Max Act: 0.6721. Min Act: -0.0001

Data Index: 373618 (C4 (Web Text))

Max Activating Token Index: 966