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

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

Neuron 2560 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9166. Min Act: -0.0001

Data Index: 1282167 (C4 (Web Text))

Max Activating Token Index: 1009

Text #1

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8456. Min Act: -0.0001

Data Index: 440308 (C4 (Web Text))

Max Activating Token Index: 698

Text #2

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9090. Min Act: -0.0001

Data Index: 854480 (C4 (Web Text))

Max Activating Token Index: 455

Text #3

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9552. Min Act: -0.0001

Data Index: 1123987 (C4 (Web Text))

Max Activating Token Index: 53

Text #4

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9664. Min Act: -0.0001

Data Index: 979412 (C4 (Web Text))

Max Activating Token Index: 187

Text #5

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8933. Min Act: -0.0001

Data Index: 147372 (C4 (Web Text))

Max Activating Token Index: 120

Text #6

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9086. Min Act: -0.0001

Data Index: 951186 (C4 (Web Text))

Max Activating Token Index: 847

Text #7

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.9263. Min Act: -0.0001

Data Index: 814541 (C4 (Web Text))

Max Activating Token Index: 686

Text #8

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8723. Min Act: -0.0001

Data Index: 141226 (C4 (Web Text))

Max Activating Token Index: 426

Text #9

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8915. Min Act: -0.0001

Data Index: 411918 (C4 (Web Text))

Max Activating Token Index: 236

Text #10

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8538. Min Act: -0.0001

Data Index: 249810 (C4 (Web Text))

Max Activating Token Index: 26

Text #11

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8301. Min Act: -0.0001

Data Index: 617404 (C4 (Web Text))

Max Activating Token Index: 547

Text #12

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.7995. Min Act: -0.0001

Data Index: 958524 (C4 (Web Text))

Max Activating Token Index: 1018

Text #13

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8479. Min Act: -0.0001

Data Index: 439081 (C4 (Web Text))

Max Activating Token Index: 802

Text #14

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8676. Min Act: -0.0001

Data Index: 445896 (C4 (Web Text))

Max Activating Token Index: 583

Text #15

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8699. Min Act: -0.0001

Data Index: 886641 (C4 (Web Text))

Max Activating Token Index: 685

Text #16

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.7944. Min Act: -0.0001

Data Index: 1650061 (Python Code)

Max Activating Token Index: 730

Text #17

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8488. Min Act: -0.0001

Data Index: 439170 (C4 (Web Text))

Max Activating Token Index: 857

Text #18

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.7820. Min Act: -0.0001

Data Index: 297490 (C4 (Web Text))

Max Activating Token Index: 473

Text #19

Max Range: 0.9664. Min Range: -0.9664

Max Act: 0.8568. Min Act: -0.0001

Data Index: 247552 (C4 (Web Text))

Max Activating Token Index: 618