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

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

Neuron 2476 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6179. Min Act: -0.0002

Data Index: 1541537 (Python Code)

Max Activating Token Index: 7

Text #1

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6188. Min Act: -0.0002

Data Index: 1030894 (C4 (Web Text))

Max Activating Token Index: 132

Text #2

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6544. Min Act: -0.0002

Data Index: 574924 (C4 (Web Text))

Max Activating Token Index: 51

Text #3

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6429. Min Act: -0.0002

Data Index: 872914 (C4 (Web Text))

Max Activating Token Index: 655

Text #4

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6202. Min Act: -0.0002

Data Index: 1120482 (C4 (Web Text))

Max Activating Token Index: 602

Text #5

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5849. Min Act: -0.0002

Data Index: 147380 (C4 (Web Text))

Max Activating Token Index: 207

Text #6

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5869. Min Act: -0.0002

Data Index: 373036 (C4 (Web Text))

Max Activating Token Index: 466

Text #7

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5965. Min Act: -0.0002

Data Index: 890201 (C4 (Web Text))

Max Activating Token Index: 703

Text #8

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6071. Min Act: -0.0002

Data Index: 1572730 (Python Code)

Max Activating Token Index: 13

Text #9

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5836. Min Act: -0.0002

Data Index: 427120 (C4 (Web Text))

Max Activating Token Index: 879

Text #10

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.6059. Min Act: -0.0002

Data Index: 511485 (C4 (Web Text))

Max Activating Token Index: 658

Text #11

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5832. Min Act: -0.0002

Data Index: 325336 (C4 (Web Text))

Max Activating Token Index: 133

Text #12

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5911. Min Act: -0.0002

Data Index: 1178919 (C4 (Web Text))

Max Activating Token Index: 96

Text #13

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5301. Min Act: -0.0002

Data Index: 1542642 (Python Code)

Max Activating Token Index: 22

Text #14

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.4481. Min Act: -0.0002

Data Index: 1571471 (Python Code)

Max Activating Token Index: 249

Text #15

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5490. Min Act: -0.0002

Data Index: 1601189 (Python Code)

Max Activating Token Index: 12

Text #16

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5490. Min Act: -0.0002

Data Index: 1439070 (Python Code)

Max Activating Token Index: 12

Text #17

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5687. Min Act: -0.0002

Data Index: 207582 (C4 (Web Text))

Max Activating Token Index: 104

Text #18

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5297. Min Act: -0.0002

Data Index: 1600333 (Python Code)

Max Activating Token Index: 5

Text #19

Max Range: 0.6544. Min Range: -0.6544

Max Act: 0.5297. Min Act: -0.0002

Data Index: 1511344 (Python Code)

Max Activating Token Index: 5