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

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

Neuron 3580 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0803. Min Act: -0.0001

Data Index: 1170063 (C4 (Web Text))

Max Activating Token Index: 540

Text #1

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0620. Min Act: -0.0001

Data Index: 1186979 (C4 (Web Text))

Max Activating Token Index: 327

Text #2

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0505. Min Act: -0.0001

Data Index: 589589 (C4 (Web Text))

Max Activating Token Index: 987

Text #3

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0528. Min Act: -0.0001

Data Index: 780902 (C4 (Web Text))

Max Activating Token Index: 660

Text #4

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0486. Min Act: -0.0001

Data Index: 232700 (C4 (Web Text))

Max Activating Token Index: 200

Text #5

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0496. Min Act: -0.0001

Data Index: 1618384 (Python Code)

Max Activating Token Index: 19

Text #6

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0510. Min Act: -0.0001

Data Index: 1501772 (Python Code)

Max Activating Token Index: 80

Text #7

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0488. Min Act: -0.0001

Data Index: 1008 (C4 (Web Text))

Max Activating Token Index: 850

Text #8

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0464. Min Act: -0.0001

Data Index: 796002 (C4 (Web Text))

Max Activating Token Index: 238

Text #9

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0468. Min Act: -0.0001

Data Index: 293140 (C4 (Web Text))

Max Activating Token Index: 529

Text #10

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0495. Min Act: -0.0001

Data Index: 1069485 (C4 (Web Text))

Max Activating Token Index: 471

Text #11

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0428. Min Act: -0.0001

Data Index: 281947 (C4 (Web Text))

Max Activating Token Index: 457

Text #12

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0458. Min Act: -0.0001

Data Index: 732055 (C4 (Web Text))

Max Activating Token Index: 292

Text #13

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0460. Min Act: -0.0001

Data Index: 1110680 (C4 (Web Text))

Max Activating Token Index: 504

Text #14

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0461. Min Act: -0.0001

Data Index: 372133 (C4 (Web Text))

Max Activating Token Index: 165

Text #15

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0441. Min Act: -0.0001

Data Index: 839810 (C4 (Web Text))

Max Activating Token Index: 145

Text #16

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0425. Min Act: -0.0001

Data Index: 431851 (C4 (Web Text))

Max Activating Token Index: 490

Text #17

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0439. Min Act: -0.0001

Data Index: 1134536 (C4 (Web Text))

Max Activating Token Index: 129

Text #18

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0406. Min Act: -0.0001

Data Index: 68932 (C4 (Web Text))

Max Activating Token Index: 328

Text #19

Max Range: 0.0803. Min Range: -0.0803

Max Act: 0.0416. Min Act: -0.0001

Data Index: 1384555 (Python Code)

Max Activating Token Index: 147