Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 1773 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.8965. Min Act: -0.0001

Data Index: 455017 (C4 (Web Text))

Max Activating Token Index: 706

Text #1

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7698. Min Act: -0.0000

Data Index: 356179 (C4 (Web Text))

Max Activating Token Index: 855

Text #2

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7301. Min Act: -0.0000

Data Index: 708451 (C4 (Web Text))

Max Activating Token Index: 867

Text #3

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7160. Min Act: -0.0000

Data Index: 636241 (C4 (Web Text))

Max Activating Token Index: 999

Text #4

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7568. Min Act: -0.0000

Data Index: 1045547 (C4 (Web Text))

Max Activating Token Index: 533

Text #5

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7441. Min Act: -0.0000

Data Index: 1082557 (C4 (Web Text))

Max Activating Token Index: 84

Text #6

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.7057. Min Act: -0.0001

Data Index: 997938 (C4 (Web Text))

Max Activating Token Index: 769

Text #7

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.6710. Min Act: -0.0001

Data Index: 793258 (C4 (Web Text))

Max Activating Token Index: 964

Text #8

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.6242. Min Act: -0.0001

Data Index: 1377797 (Python Code)

Max Activating Token Index: 19

Text #9

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.6239. Min Act: -0.0001

Data Index: 1536197 (Python Code)

Max Activating Token Index: 26

Text #10

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5604. Min Act: -0.0000

Data Index: 514332 (C4 (Web Text))

Max Activating Token Index: 391

Text #11

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5505. Min Act: -0.0001

Data Index: 1554104 (Python Code)

Max Activating Token Index: 23

Text #12

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5664. Min Act: -0.0000

Data Index: 825476 (C4 (Web Text))

Max Activating Token Index: 934

Text #13

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5813. Min Act: -0.0001

Data Index: 1584038 (Python Code)

Max Activating Token Index: 23

Text #14

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5813. Min Act: -0.0001

Data Index: 1393229 (Python Code)

Max Activating Token Index: 23

Text #15

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5894. Min Act: -0.0001

Data Index: 442186 (C4 (Web Text))

Max Activating Token Index: 897

Text #16

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5447. Min Act: -0.0001

Data Index: 1449157 (Python Code)

Max Activating Token Index: 21

Text #17

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.4849. Min Act: -0.0001

Data Index: 1376902 (Python Code)

Max Activating Token Index: 707

Text #18

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5983. Min Act: -0.0001

Data Index: 1193291 (C4 (Web Text))

Max Activating Token Index: 239

Text #19

Max Range: 0.8965. Min Range: -0.8965

Max Act: 0.5596. Min Act: -0.0000

Data Index: 271495 (C4 (Web Text))

Max Activating Token Index: 91