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

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

Neuron 1914 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.3350. Min Range: -5.3350

Max Act: 5.3350. Min Act: -0.0000

Data Index: 55430 (C4 (Web Text))

Max Activating Token Index: 853

Text #1

Max Range: 5.3350. Min Range: -5.3350

Max Act: 4.6127. Min Act: -0.0000

Data Index: 203615 (C4 (Web Text))

Max Activating Token Index: 74

Text #2

Max Range: 5.3350. Min Range: -5.3350

Max Act: 4.4244. Min Act: -0.0000

Data Index: 132579 (C4 (Web Text))

Max Activating Token Index: 861

Text #3

Max Range: 5.3350. Min Range: -5.3350

Max Act: 4.3456. Min Act: -0.0000

Data Index: 48645 (C4 (Web Text))

Max Activating Token Index: 796

Text #4

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.8544. Min Act: -0.0000

Data Index: 1169428 (C4 (Web Text))

Max Activating Token Index: 831

Text #5

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.7952. Min Act: -0.0000

Data Index: 1630087 (Python Code)

Max Activating Token Index: 126

Text #6

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.9188. Min Act: -0.0000

Data Index: 661778 (C4 (Web Text))

Max Activating Token Index: 898

Text #7

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.8549. Min Act: -0.0000

Data Index: 700238 (C4 (Web Text))

Max Activating Token Index: 526

Text #8

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.6682. Min Act: -0.0000

Data Index: 291020 (C4 (Web Text))

Max Activating Token Index: 560

Text #9

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.3845. Min Act: -0.0000

Data Index: 1491116 (Python Code)

Max Activating Token Index: 213

Text #10

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.5051. Min Act: -0.0000

Data Index: 1057550 (C4 (Web Text))

Max Activating Token Index: 453

Text #11

Max Range: 5.3350. Min Range: -5.3350

Max Act: 2.9997. Min Act: -0.0000

Data Index: 1020802 (C4 (Web Text))

Max Activating Token Index: 376

Text #12

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.3944. Min Act: -0.0000

Data Index: 288843 (C4 (Web Text))

Max Activating Token Index: 496

Text #13

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.2426. Min Act: -0.0000

Data Index: 1522510 (Python Code)

Max Activating Token Index: 813

Text #14

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.3438. Min Act: -0.0000

Data Index: 1167699 (C4 (Web Text))

Max Activating Token Index: 320

Text #15

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.2386. Min Act: -0.0000

Data Index: 1460380 (Python Code)

Max Activating Token Index: 951

Text #16

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.0654. Min Act: -0.0000

Data Index: 1167646 (C4 (Web Text))

Max Activating Token Index: 475

Text #17

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.4987. Min Act: -0.0000

Data Index: 1389198 (Python Code)

Max Activating Token Index: 965

Text #18

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.3295. Min Act: -0.0000

Data Index: 1247851 (C4 (Web Text))

Max Activating Token Index: 596

Text #19

Max Range: 5.3350. Min Range: -5.3350

Max Act: 3.3261. Min Act: -0.0000

Data Index: 533218 (C4 (Web Text))

Max Activating Token Index: 889