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

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

Neuron 4796 in Layer 9

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6958. Min Act: -0.0000

Data Index: 880683 (C4 (Web Text))

Max Activating Token Index: 360

Text #1

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6560. Min Act: -0.0000

Data Index: 941121 (C4 (Web Text))

Max Activating Token Index: 841

Text #2

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6713. Min Act: -0.0000

Data Index: 559618 (C4 (Web Text))

Max Activating Token Index: 899

Text #3

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6091. Min Act: -0.0000

Data Index: 875247 (C4 (Web Text))

Max Activating Token Index: 358

Text #4

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6343. Min Act: -0.0000

Data Index: 513622 (C4 (Web Text))

Max Activating Token Index: 983

Text #5

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6280. Min Act: -0.0000

Data Index: 1395028 (Python Code)

Max Activating Token Index: 607

Text #6

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5812. Min Act: -0.0000

Data Index: 1354009 (C4 (Web Text))

Max Activating Token Index: 196

Text #7

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6488. Min Act: -0.0000

Data Index: 1041147 (C4 (Web Text))

Max Activating Token Index: 349

Text #8

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5743. Min Act: -0.0000

Data Index: 1052451 (C4 (Web Text))

Max Activating Token Index: 921

Text #9

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5953. Min Act: -0.0000

Data Index: 886566 (C4 (Web Text))

Max Activating Token Index: 1004

Text #10

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5939. Min Act: -0.0000

Data Index: 79619 (C4 (Web Text))

Max Activating Token Index: 950

Text #11

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.6007. Min Act: -0.0000

Data Index: 1327660 (C4 (Web Text))

Max Activating Token Index: 847

Text #12

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5655. Min Act: -0.0000

Data Index: 158007 (C4 (Web Text))

Max Activating Token Index: 870

Text #13

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5649. Min Act: -0.0000

Data Index: 1495586 (Python Code)

Max Activating Token Index: 1019

Text #14

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5620. Min Act: -0.0000

Data Index: 1493313 (Python Code)

Max Activating Token Index: 74

Text #15

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5024. Min Act: -0.0000

Data Index: 1393896 (Python Code)

Max Activating Token Index: 752

Text #16

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5660. Min Act: -0.0000

Data Index: 338012 (C4 (Web Text))

Max Activating Token Index: 635

Text #17

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5411. Min Act: -0.0000

Data Index: 921462 (C4 (Web Text))

Max Activating Token Index: 694

Text #18

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5553. Min Act: -0.0000

Data Index: 268758 (C4 (Web Text))

Max Activating Token Index: 328

Text #19

Max Range: 0.6958. Min Range: -0.6958

Max Act: 0.5666. Min Act: -0.0000

Data Index: 61657 (C4 (Web Text))

Max Activating Token Index: 352