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

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

Neuron 2514 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0534. Min Act: -0.0001

Data Index: 866268 (C4 (Web Text))

Max Activating Token Index: 31

Text #1

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0497. Min Act: -0.0001

Data Index: 476002 (C4 (Web Text))

Max Activating Token Index: 691

Text #2

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0417. Min Act: -0.0001

Data Index: 975523 (C4 (Web Text))

Max Activating Token Index: 60

Text #3

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0411. Min Act: -0.0001

Data Index: 565882 (C4 (Web Text))

Max Activating Token Index: 509

Text #4

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0382. Min Act: -0.0001

Data Index: 909299 (C4 (Web Text))

Max Activating Token Index: 14

Text #5

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0393. Min Act: -0.0001

Data Index: 99028 (C4 (Web Text))

Max Activating Token Index: 499

Text #6

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0387. Min Act: -0.0001

Data Index: 267738 (C4 (Web Text))

Max Activating Token Index: 133

Text #7

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0354. Min Act: -0.0001

Data Index: 857383 (C4 (Web Text))

Max Activating Token Index: 681

Text #8

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0348. Min Act: -0.0001

Data Index: 365193 (C4 (Web Text))

Max Activating Token Index: 500

Text #9

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0377. Min Act: -0.0001

Data Index: 1025366 (C4 (Web Text))

Max Activating Token Index: 206

Text #10

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0329. Min Act: -0.0001

Data Index: 78418 (C4 (Web Text))

Max Activating Token Index: 30

Text #11

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0314. Min Act: -0.0001

Data Index: 118124 (C4 (Web Text))

Max Activating Token Index: 368

Text #12

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0332. Min Act: -0.0001

Data Index: 1637048 (Python Code)

Max Activating Token Index: 85

Text #13

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0310. Min Act: -0.0001

Data Index: 874750 (C4 (Web Text))

Max Activating Token Index: 500

Text #14

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0331. Min Act: -0.0001

Data Index: 262695 (C4 (Web Text))

Max Activating Token Index: 65

Text #15

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0311. Min Act: -0.0001

Data Index: 126537 (C4 (Web Text))

Max Activating Token Index: 182

Text #16

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0330. Min Act: -0.0001

Data Index: 71685 (C4 (Web Text))

Max Activating Token Index: 221

Text #17

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0320. Min Act: -0.0001

Data Index: 1205652 (C4 (Web Text))

Max Activating Token Index: 132

Text #18

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0320. Min Act: -0.0001

Data Index: 37937 (C4 (Web Text))

Max Activating Token Index: 252

Text #19

Max Range: 0.0534. Min Range: -0.0534

Max Act: 0.0290. Min Act: -0.0001

Data Index: 33984 (C4 (Web Text))

Max Activating Token Index: 175