Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1789 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0417. Min Act: -0.0003

Data Index: 1532042 (Python Code)

Max Activating Token Index: 995

Text #1

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0336. Min Act: -0.0003

Data Index: 272507 (C4 (Web Text))

Max Activating Token Index: 754

Text #2

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0335. Min Act: -0.0002

Data Index: 105550 (C4 (Web Text))

Max Activating Token Index: 947

Text #3

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0309. Min Act: -0.0002

Data Index: 920473 (C4 (Web Text))

Max Activating Token Index: 941

Text #4

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0295. Min Act: -0.0003

Data Index: 1363315 (Python Code)

Max Activating Token Index: 225

Text #5

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0303. Min Act: -0.0004

Data Index: 305480 (C4 (Web Text))

Max Activating Token Index: 654

Text #6

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0306. Min Act: -0.0003

Data Index: 456607 (C4 (Web Text))

Max Activating Token Index: 318

Text #7

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0300. Min Act: -0.0003

Data Index: 226021 (C4 (Web Text))

Max Activating Token Index: 640

Text #8

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0290. Min Act: -0.0003

Data Index: 1642745 (Python Code)

Max Activating Token Index: 218

Text #9

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0300. Min Act: -0.0003

Data Index: 1428755 (Python Code)

Max Activating Token Index: 311

Text #10

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0275. Min Act: -0.0003

Data Index: 1015978 (C4 (Web Text))

Max Activating Token Index: 925

Text #11

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0292. Min Act: -0.0003

Data Index: 1181574 (C4 (Web Text))

Max Activating Token Index: 905

Text #12

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0270. Min Act: -0.0002

Data Index: 839508 (C4 (Web Text))

Max Activating Token Index: 920

Text #13

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0277. Min Act: -0.0003

Data Index: 53555 (C4 (Web Text))

Max Activating Token Index: 268

Text #14

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0266. Min Act: -0.0004

Data Index: 1305549 (C4 (Web Text))

Max Activating Token Index: 699

Text #15

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0296. Min Act: -0.0003

Data Index: 1104296 (C4 (Web Text))

Max Activating Token Index: 634

Text #16

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0280. Min Act: -0.0003

Data Index: 445084 (C4 (Web Text))

Max Activating Token Index: 840

Text #17

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0274. Min Act: -0.0003

Data Index: 1640943 (Python Code)

Max Activating Token Index: 203

Text #18

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0272. Min Act: -0.0002

Data Index: 165744 (C4 (Web Text))

Max Activating Token Index: 137

Text #19

Max Range: 0.0417. Min Range: -0.0417

Max Act: 0.0287. Min Act: -0.0003

Data Index: 701815 (C4 (Web Text))

Max Activating Token Index: 973