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

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

Neuron 648 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.3452. Min Act: -0.0000

Data Index: 848667 (C4 (Web Text))

Max Activating Token Index: 852

Text #1

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.3115. Min Act: -0.0000

Data Index: 1030648 (C4 (Web Text))

Max Activating Token Index: 1007

Text #2

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.3209. Min Act: -0.0001

Data Index: 746775 (C4 (Web Text))

Max Activating Token Index: 439

Text #3

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2736. Min Act: -0.0000

Data Index: 17479 (C4 (Web Text))

Max Activating Token Index: 181

Text #4

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2733. Min Act: -0.0001

Data Index: 664687 (C4 (Web Text))

Max Activating Token Index: 655

Text #5

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2678. Min Act: -0.0000

Data Index: 1424052 (Python Code)

Max Activating Token Index: 57

Text #6

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2656. Min Act: -0.0000

Data Index: 872858 (C4 (Web Text))

Max Activating Token Index: 390

Text #7

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2636. Min Act: -0.0000

Data Index: 542897 (C4 (Web Text))

Max Activating Token Index: 473

Text #8

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2630. Min Act: -0.0000

Data Index: 304362 (C4 (Web Text))

Max Activating Token Index: 1009

Text #9

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2478. Min Act: -0.0000

Data Index: 146676 (C4 (Web Text))

Max Activating Token Index: 673

Text #10

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2537. Min Act: -0.0000

Data Index: 28981 (C4 (Web Text))

Max Activating Token Index: 148

Text #11

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2387. Min Act: -0.0001

Data Index: 901666 (C4 (Web Text))

Max Activating Token Index: 343

Text #12

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2416. Min Act: -0.0000

Data Index: 928093 (C4 (Web Text))

Max Activating Token Index: 436

Text #13

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2444. Min Act: -0.0000

Data Index: 754456 (C4 (Web Text))

Max Activating Token Index: 101

Text #14

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2520. Min Act: -0.0000

Data Index: 1448533 (Python Code)

Max Activating Token Index: 306

Text #15

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2289. Min Act: -0.0001

Data Index: 739402 (C4 (Web Text))

Max Activating Token Index: 180

Text #16

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2342. Min Act: -0.0000

Data Index: 989345 (C4 (Web Text))

Max Activating Token Index: 601

Text #17

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2484. Min Act: -0.0000

Data Index: 707163 (C4 (Web Text))

Max Activating Token Index: 516

Text #18

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2333. Min Act: -0.0001

Data Index: 1355463 (C4 (Web Text))

Max Activating Token Index: 374

Text #19

Max Range: 0.3452. Min Range: -0.3452

Max Act: 0.2156. Min Act: -0.0000

Data Index: 1140449 (C4 (Web Text))

Max Activating Token Index: 160