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

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

Neuron 340 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0542. Min Act: -0.0006

Data Index: 921503 (C4 (Web Text))

Max Activating Token Index: 92

Text #1

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0482. Min Act: -0.0006

Data Index: 168475 (C4 (Web Text))

Max Activating Token Index: 886

Text #2

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0499. Min Act: -0.0006

Data Index: 841603 (C4 (Web Text))

Max Activating Token Index: 49

Text #3

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0490. Min Act: -0.0006

Data Index: 250720 (C4 (Web Text))

Max Activating Token Index: 52

Text #4

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0458. Min Act: -0.0006

Data Index: 1275731 (C4 (Web Text))

Max Activating Token Index: 866

Text #5

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0481. Min Act: -0.0006

Data Index: 552157 (C4 (Web Text))

Max Activating Token Index: 591

Text #6

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0439. Min Act: -0.0005

Data Index: 801071 (C4 (Web Text))

Max Activating Token Index: 83

Text #7

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0436. Min Act: -0.0006

Data Index: 35093 (C4 (Web Text))

Max Activating Token Index: 1021

Text #8

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0444. Min Act: -0.0006

Data Index: 1114751 (C4 (Web Text))

Max Activating Token Index: 28

Text #9

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0444. Min Act: -0.0006

Data Index: 667538 (C4 (Web Text))

Max Activating Token Index: 25

Text #10

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0422. Min Act: -0.0006

Data Index: 453025 (C4 (Web Text))

Max Activating Token Index: 971

Text #11

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0444. Min Act: -0.0006

Data Index: 675825 (C4 (Web Text))

Max Activating Token Index: 979

Text #12

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0434. Min Act: -0.0005

Data Index: 485156 (C4 (Web Text))

Max Activating Token Index: 139

Text #13

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0397. Min Act: -0.0006

Data Index: 910885 (C4 (Web Text))

Max Activating Token Index: 1014

Text #14

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0412. Min Act: -0.0006

Data Index: 1592415 (Python Code)

Max Activating Token Index: 639

Text #15

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0419. Min Act: -0.0005

Data Index: 1258813 (C4 (Web Text))

Max Activating Token Index: 914

Text #16

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0407. Min Act: -0.0005

Data Index: 1118853 (C4 (Web Text))

Max Activating Token Index: 784

Text #17

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0429. Min Act: -0.0005

Data Index: 127355 (C4 (Web Text))

Max Activating Token Index: 687

Text #18

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0410. Min Act: -0.0006

Data Index: 1077996 (C4 (Web Text))

Max Activating Token Index: 362

Text #19

Max Range: 0.0542. Min Range: -0.0542

Max Act: 0.0414. Min Act: -0.0006

Data Index: 1303599 (C4 (Web Text))

Max Activating Token Index: 123