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

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

Neuron 158 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: 1.1952. Min Range: -1.1952

Max Act: 1.1952. Min Act: -0.0001

Data Index: 568539 (C4 (Web Text))

Max Activating Token Index: 80

Text #1

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.1357. Min Act: -0.0001

Data Index: 1130352 (C4 (Web Text))

Max Activating Token Index: 152

Text #2

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0891. Min Act: -0.0001

Data Index: 1384946 (Python Code)

Max Activating Token Index: 660

Text #3

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0690. Min Act: -0.0001

Data Index: 928771 (C4 (Web Text))

Max Activating Token Index: 719

Text #4

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0447. Min Act: -0.0001

Data Index: 1063472 (C4 (Web Text))

Max Activating Token Index: 177

Text #5

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9748. Min Act: -0.0001

Data Index: 1478328 (Python Code)

Max Activating Token Index: 212

Text #6

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0547. Min Act: -0.0001

Data Index: 192478 (C4 (Web Text))

Max Activating Token Index: 10

Text #7

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9655. Min Act: -0.0001

Data Index: 464340 (C4 (Web Text))

Max Activating Token Index: 56

Text #8

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9419. Min Act: -0.0001

Data Index: 1625236 (Python Code)

Max Activating Token Index: 101

Text #9

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0257. Min Act: -0.0001

Data Index: 20223 (C4 (Web Text))

Max Activating Token Index: 561

Text #10

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.8929. Min Act: -0.0001

Data Index: 391331 (C4 (Web Text))

Max Activating Token Index: 306

Text #11

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9645. Min Act: -0.0001

Data Index: 1106850 (C4 (Web Text))

Max Activating Token Index: 670

Text #12

Max Range: 1.1952. Min Range: -1.1952

Max Act: 1.0406. Min Act: -0.0001

Data Index: 905622 (C4 (Web Text))

Max Activating Token Index: 205

Text #13

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9508. Min Act: -0.0001

Data Index: 693679 (C4 (Web Text))

Max Activating Token Index: 255

Text #14

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9806. Min Act: -0.0001

Data Index: 1460049 (Python Code)

Max Activating Token Index: 435

Text #15

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.8782. Min Act: -0.0001

Data Index: 1333674 (C4 (Web Text))

Max Activating Token Index: 102

Text #16

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.8939. Min Act: -0.0001

Data Index: 1532575 (Python Code)

Max Activating Token Index: 425

Text #17

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.8709. Min Act: -0.0001

Data Index: 664836 (C4 (Web Text))

Max Activating Token Index: 288

Text #18

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.9222. Min Act: -0.0001

Data Index: 526100 (C4 (Web Text))

Max Activating Token Index: 247

Text #19

Max Range: 1.1952. Min Range: -1.1952

Max Act: 0.8395. Min Act: -0.0001

Data Index: 372228 (C4 (Web Text))

Max Activating Token Index: 248