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

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

Neuron 4096 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: 0.1408. Min Range: -0.1408

Max Act: 0.1343. Min Act: -0.0000

Data Index: 721339 (C4 (Web Text))

Max Activating Token Index: 724

Text #1

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1408. Min Act: -0.0000

Data Index: 743533 (C4 (Web Text))

Max Activating Token Index: 320

Text #2

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1261. Min Act: -0.0000

Data Index: 473463 (C4 (Web Text))

Max Activating Token Index: 231

Text #3

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1237. Min Act: -0.0000

Data Index: 573658 (C4 (Web Text))

Max Activating Token Index: 982

Text #4

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1188. Min Act: -0.0001

Data Index: 1334705 (C4 (Web Text))

Max Activating Token Index: 159

Text #5

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1181. Min Act: -0.0000

Data Index: 420589 (C4 (Web Text))

Max Activating Token Index: 420

Text #6

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1172. Min Act: -0.0000

Data Index: 7487 (C4 (Web Text))

Max Activating Token Index: 637

Text #7

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1115. Min Act: -0.0001

Data Index: 383044 (C4 (Web Text))

Max Activating Token Index: 200

Text #8

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1120. Min Act: -0.0001

Data Index: 332543 (C4 (Web Text))

Max Activating Token Index: 289

Text #9

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1191. Min Act: -0.0001

Data Index: 803221 (C4 (Web Text))

Max Activating Token Index: 279

Text #10

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1151. Min Act: -0.0001

Data Index: 876885 (C4 (Web Text))

Max Activating Token Index: 997

Text #11

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1220. Min Act: -0.0001

Data Index: 803863 (C4 (Web Text))

Max Activating Token Index: 533

Text #12

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1158. Min Act: -0.0001

Data Index: 869425 (C4 (Web Text))

Max Activating Token Index: 737

Text #13

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.0979. Min Act: -0.0001

Data Index: 1424838 (Python Code)

Max Activating Token Index: 705

Text #14

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1049. Min Act: -0.0000

Data Index: 179452 (C4 (Web Text))

Max Activating Token Index: 879

Text #15

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.0965. Min Act: -0.0000

Data Index: 716060 (C4 (Web Text))

Max Activating Token Index: 166

Text #16

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1060. Min Act: -0.0000

Data Index: 1559019 (Python Code)

Max Activating Token Index: 448

Text #17

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1086. Min Act: -0.0000

Data Index: 1087217 (C4 (Web Text))

Max Activating Token Index: 195

Text #18

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.1033. Min Act: -0.0000

Data Index: 823501 (C4 (Web Text))

Max Activating Token Index: 514

Text #19

Max Range: 0.1408. Min Range: -0.1408

Max Act: 0.0959. Min Act: -0.0000

Data Index: 577132 (C4 (Web Text))

Max Activating Token Index: 1010