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

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

Neuron 273 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4882. Min Act: -0.0001

Data Index: 659715 (C4 (Web Text))

Max Activating Token Index: 446

Text #1

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4430. Min Act: -0.0001

Data Index: 187933 (C4 (Web Text))

Max Activating Token Index: 620

Text #2

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4480. Min Act: -0.0001

Data Index: 253434 (C4 (Web Text))

Max Activating Token Index: 582

Text #3

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4403. Min Act: -0.0001

Data Index: 59788 (C4 (Web Text))

Max Activating Token Index: 90

Text #4

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4600. Min Act: -0.0001

Data Index: 434680 (C4 (Web Text))

Max Activating Token Index: 771

Text #5

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4321. Min Act: -0.0001

Data Index: 1356968 (C4 (Web Text))

Max Activating Token Index: 921

Text #6

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4317. Min Act: -0.0001

Data Index: 1302422 (C4 (Web Text))

Max Activating Token Index: 821

Text #7

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4343. Min Act: -0.0001

Data Index: 204010 (C4 (Web Text))

Max Activating Token Index: 832

Text #8

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4335. Min Act: -0.0001

Data Index: 35534 (C4 (Web Text))

Max Activating Token Index: 665

Text #9

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4217. Min Act: -0.0001

Data Index: 831147 (C4 (Web Text))

Max Activating Token Index: 257

Text #10

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4451. Min Act: -0.0001

Data Index: 1125702 (C4 (Web Text))

Max Activating Token Index: 95

Text #11

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4083. Min Act: -0.0001

Data Index: 540132 (C4 (Web Text))

Max Activating Token Index: 375

Text #12

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4071. Min Act: -0.0001

Data Index: 109254 (C4 (Web Text))

Max Activating Token Index: 737

Text #13

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4196. Min Act: -0.0001

Data Index: 446149 (C4 (Web Text))

Max Activating Token Index: 748

Text #14

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4173. Min Act: -0.0001

Data Index: 1223578 (C4 (Web Text))

Max Activating Token Index: 746

Text #15

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4157. Min Act: -0.0001

Data Index: 629317 (C4 (Web Text))

Max Activating Token Index: 964

Text #16

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.3960. Min Act: -0.0001

Data Index: 895516 (C4 (Web Text))

Max Activating Token Index: 540

Text #17

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4010. Min Act: -0.0001

Data Index: 445108 (C4 (Web Text))

Max Activating Token Index: 887

Text #18

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.4118. Min Act: -0.0001

Data Index: 321502 (C4 (Web Text))

Max Activating Token Index: 87

Text #19

Max Range: 0.4882. Min Range: -0.4882

Max Act: 0.3868. Min Act: -0.0001

Data Index: 89976 (C4 (Web Text))

Max Activating Token Index: 775