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

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

Neuron 4325 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: 0.3296. Min Range: -0.3296

Max Act: 0.3081. Min Act: -0.0001

Data Index: 911963 (C4 (Web Text))

Max Activating Token Index: 614

Text #1

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.3131. Min Act: -0.0001

Data Index: 1459016 (Python Code)

Max Activating Token Index: 407

Text #2

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.3296. Min Act: -0.0001

Data Index: 1018516 (C4 (Web Text))

Max Activating Token Index: 800

Text #3

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.3012. Min Act: -0.0001

Data Index: 119211 (C4 (Web Text))

Max Activating Token Index: 477

Text #4

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2787. Min Act: -0.0001

Data Index: 993856 (C4 (Web Text))

Max Activating Token Index: 536

Text #5

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2669. Min Act: -0.0001

Data Index: 347954 (C4 (Web Text))

Max Activating Token Index: 983

Text #6

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2762. Min Act: -0.0001

Data Index: 927971 (C4 (Web Text))

Max Activating Token Index: 386

Text #7

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2766. Min Act: -0.0001

Data Index: 1244471 (C4 (Web Text))

Max Activating Token Index: 805

Text #8

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2632. Min Act: -0.0001

Data Index: 528492 (C4 (Web Text))

Max Activating Token Index: 605

Text #9

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2581. Min Act: -0.0001

Data Index: 294279 (C4 (Web Text))

Max Activating Token Index: 174

Text #10

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2381. Min Act: -0.0001

Data Index: 825300 (C4 (Web Text))

Max Activating Token Index: 331

Text #11

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2733. Min Act: -0.0001

Data Index: 787262 (C4 (Web Text))

Max Activating Token Index: 630

Text #12

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2738. Min Act: -0.0001

Data Index: 206265 (C4 (Web Text))

Max Activating Token Index: 754

Text #13

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2604. Min Act: -0.0001

Data Index: 1435124 (Python Code)

Max Activating Token Index: 445

Text #14

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2655. Min Act: -0.0001

Data Index: 1219531 (C4 (Web Text))

Max Activating Token Index: 519

Text #15

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2551. Min Act: -0.0001

Data Index: 361348 (C4 (Web Text))

Max Activating Token Index: 5

Text #16

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2465. Min Act: -0.0001

Data Index: 980907 (C4 (Web Text))

Max Activating Token Index: 229

Text #17

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2645. Min Act: -0.0001

Data Index: 865150 (C4 (Web Text))

Max Activating Token Index: 810

Text #18

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2596. Min Act: -0.0001

Data Index: 1568711 (Python Code)

Max Activating Token Index: 142

Text #19

Max Range: 0.3296. Min Range: -0.3296

Max Act: 0.2515. Min Act: -0.0001

Data Index: 1352862 (C4 (Web Text))

Max Activating Token Index: 372