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

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

Neuron 1323 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: 0.7561. Min Range: -0.7561

Max Act: 0.7561. Min Act: -0.0001

Data Index: 233805 (C4 (Web Text))

Max Activating Token Index: 506

Text #1

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.7045. Min Act: -0.0001

Data Index: 537530 (C4 (Web Text))

Max Activating Token Index: 747

Text #2

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.7312. Min Act: -0.0001

Data Index: 892196 (C4 (Web Text))

Max Activating Token Index: 526

Text #3

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6999. Min Act: -0.0001

Data Index: 70115 (C4 (Web Text))

Max Activating Token Index: 501

Text #4

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.7441. Min Act: -0.0001

Data Index: 1123909 (C4 (Web Text))

Max Activating Token Index: 557

Text #5

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6429. Min Act: -0.0001

Data Index: 85175 (C4 (Web Text))

Max Activating Token Index: 28

Text #6

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6245. Min Act: -0.0001

Data Index: 722151 (C4 (Web Text))

Max Activating Token Index: 319

Text #7

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6469. Min Act: -0.0001

Data Index: 750009 (C4 (Web Text))

Max Activating Token Index: 823

Text #8

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6749. Min Act: -0.0001

Data Index: 1201757 (C4 (Web Text))

Max Activating Token Index: 28

Text #9

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6559. Min Act: -0.0001

Data Index: 1564228 (Python Code)

Max Activating Token Index: 5

Text #10

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6177. Min Act: -0.0001

Data Index: 1235620 (C4 (Web Text))

Max Activating Token Index: 920

Text #11

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6097. Min Act: -0.0001

Data Index: 1608326 (Python Code)

Max Activating Token Index: 468

Text #12

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.5906. Min Act: -0.0001

Data Index: 950089 (C4 (Web Text))

Max Activating Token Index: 275

Text #13

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6224. Min Act: -0.0001

Data Index: 1503491 (Python Code)

Max Activating Token Index: 8

Text #14

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6220. Min Act: -0.0001

Data Index: 1290208 (C4 (Web Text))

Max Activating Token Index: 664

Text #15

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6342. Min Act: -0.0001

Data Index: 119051 (C4 (Web Text))

Max Activating Token Index: 224

Text #16

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.5516. Min Act: -0.0001

Data Index: 957681 (C4 (Web Text))

Max Activating Token Index: 531

Text #17

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.5374. Min Act: -0.0001

Data Index: 86591 (C4 (Web Text))

Max Activating Token Index: 875

Text #18

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6054. Min Act: -0.0001

Data Index: 1186619 (C4 (Web Text))

Max Activating Token Index: 440

Text #19

Max Range: 0.7561. Min Range: -0.7561

Max Act: 0.6393. Min Act: -0.0001

Data Index: 290554 (C4 (Web Text))

Max Activating Token Index: 115