Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 1781 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.6273. Min Act: -0.0001

Data Index: 368838 (C4 (Web Text))

Max Activating Token Index: 69

Text #1

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.7377. Min Act: -0.0001

Data Index: 895339 (C4 (Web Text))

Max Activating Token Index: 388

Text #2

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.7196. Min Act: -0.0001

Data Index: 818027 (C4 (Web Text))

Max Activating Token Index: 963

Text #3

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.8839. Min Act: -0.0001

Data Index: 833115 (C4 (Web Text))

Max Activating Token Index: 285

Text #4

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.6814. Min Act: -0.0001

Data Index: 1558981 (Python Code)

Max Activating Token Index: 23

Text #5

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.6470. Min Act: -0.0001

Data Index: 193072 (C4 (Web Text))

Max Activating Token Index: 953

Text #6

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.6718. Min Act: -0.0001

Data Index: 203196 (C4 (Web Text))

Max Activating Token Index: 557

Text #7

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.5333. Min Act: -0.0001

Data Index: 1103863 (C4 (Web Text))

Max Activating Token Index: 309

Text #8

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.8048. Min Act: -0.0001

Data Index: 299577 (C4 (Web Text))

Max Activating Token Index: 267

Text #9

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4295. Min Act: -0.0001

Data Index: 754305 (C4 (Web Text))

Max Activating Token Index: 124

Text #10

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4192. Min Act: -0.0001

Data Index: 204963 (C4 (Web Text))

Max Activating Token Index: 310

Text #11

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.5664. Min Act: -0.0001

Data Index: 488606 (C4 (Web Text))

Max Activating Token Index: 363

Text #12

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4857. Min Act: -0.0001

Data Index: 1099554 (C4 (Web Text))

Max Activating Token Index: 911

Text #13

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4911. Min Act: -0.0001

Data Index: 1410336 (Python Code)

Max Activating Token Index: 293

Text #14

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4843. Min Act: -0.0001

Data Index: 1444759 (Python Code)

Max Activating Token Index: 72

Text #15

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4004. Min Act: -0.0001

Data Index: 1278508 (C4 (Web Text))

Max Activating Token Index: 594

Text #16

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4170. Min Act: -0.0001

Data Index: 902615 (C4 (Web Text))

Max Activating Token Index: 175

Text #17

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.3836. Min Act: -0.0001

Data Index: 1231207 (C4 (Web Text))

Max Activating Token Index: 296

Text #18

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.6678. Min Act: -0.0001

Data Index: 1000942 (C4 (Web Text))

Max Activating Token Index: 652

Text #19

Max Range: 1.8839. Min Range: -1.8839

Max Act: 1.4745. Min Act: -0.0001

Data Index: 1559349 (Python Code)

Max Activating Token Index: 797