Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 515 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0529. Min Act: -0.0004

Data Index: 1245729 (C4 (Web Text))

Max Activating Token Index: 871

Text #1

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0494. Min Act: -0.0003

Data Index: 711199 (C4 (Web Text))

Max Activating Token Index: 1018

Text #2

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0493. Min Act: -0.0003

Data Index: 1019577 (C4 (Web Text))

Max Activating Token Index: 657

Text #3

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0464. Min Act: -0.0002

Data Index: 148871 (C4 (Web Text))

Max Activating Token Index: 176

Text #4

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0454. Min Act: -0.0003

Data Index: 907765 (C4 (Web Text))

Max Activating Token Index: 285

Text #5

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0461. Min Act: -0.0004

Data Index: 10807 (C4 (Web Text))

Max Activating Token Index: 454

Text #6

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0469. Min Act: -0.0003

Data Index: 109740 (C4 (Web Text))

Max Activating Token Index: 778

Text #7

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0442. Min Act: -0.0003

Data Index: 190999 (C4 (Web Text))

Max Activating Token Index: 263

Text #8

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0448. Min Act: -0.0003

Data Index: 908550 (C4 (Web Text))

Max Activating Token Index: 535

Text #9

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0452. Min Act: -0.0003

Data Index: 154266 (C4 (Web Text))

Max Activating Token Index: 241

Text #10

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0456. Min Act: -0.0003

Data Index: 654191 (C4 (Web Text))

Max Activating Token Index: 654

Text #11

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0444. Min Act: -0.0003

Data Index: 1597056 (Python Code)

Max Activating Token Index: 382

Text #12

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0451. Min Act: -0.0004

Data Index: 666476 (C4 (Web Text))

Max Activating Token Index: 805

Text #13

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0428. Min Act: -0.0003

Data Index: 521946 (C4 (Web Text))

Max Activating Token Index: 792

Text #14

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0426. Min Act: -0.0003

Data Index: 1069514 (C4 (Web Text))

Max Activating Token Index: 617

Text #15

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0423. Min Act: -0.0003

Data Index: 1435272 (Python Code)

Max Activating Token Index: 536

Text #16

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0440. Min Act: -0.0003

Data Index: 320319 (C4 (Web Text))

Max Activating Token Index: 684

Text #17

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0419. Min Act: -0.0003

Data Index: 793277 (C4 (Web Text))

Max Activating Token Index: 55

Text #18

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0427. Min Act: -0.0003

Data Index: 8633 (C4 (Web Text))

Max Activating Token Index: 765

Text #19

Max Range: 0.0529. Min Range: -0.0529

Max Act: 0.0425. Min Act: -0.0003

Data Index: 108239 (C4 (Web Text))

Max Activating Token Index: 82