Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 98 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1302. Min Act: -0.0001

Data Index: 890870 (C4 (Web Text))

Max Activating Token Index: 370

Text #1

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1074. Min Act: -0.0001

Data Index: 13042 (C4 (Web Text))

Max Activating Token Index: 299

Text #2

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1477. Min Act: -0.0001

Data Index: 100718 (C4 (Web Text))

Max Activating Token Index: 213

Text #3

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1212. Min Act: -0.0001

Data Index: 1611128 (Python Code)

Max Activating Token Index: 185

Text #4

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1286. Min Act: -0.0001

Data Index: 535178 (C4 (Web Text))

Max Activating Token Index: 595

Text #5

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1169. Min Act: -0.0001

Data Index: 1184615 (C4 (Web Text))

Max Activating Token Index: 210

Text #6

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1303. Min Act: -0.0001

Data Index: 489560 (C4 (Web Text))

Max Activating Token Index: 15

Text #7

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1351. Min Act: -0.0001

Data Index: 813482 (C4 (Web Text))

Max Activating Token Index: 425

Text #8

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1182. Min Act: -0.0001

Data Index: 226259 (C4 (Web Text))

Max Activating Token Index: 349

Text #9

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1360. Min Act: -0.0001

Data Index: 579232 (C4 (Web Text))

Max Activating Token Index: 285

Text #10

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1322. Min Act: -0.0001

Data Index: 968277 (C4 (Web Text))

Max Activating Token Index: 37

Text #11

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1208. Min Act: -0.0001

Data Index: 1487743 (Python Code)

Max Activating Token Index: 21

Text #12

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.0952. Min Act: -0.0001

Data Index: 786571 (C4 (Web Text))

Max Activating Token Index: 832

Text #13

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1375. Min Act: -0.0001

Data Index: 130891 (C4 (Web Text))

Max Activating Token Index: 564

Text #14

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1334. Min Act: -0.0001

Data Index: 33320 (C4 (Web Text))

Max Activating Token Index: 488

Text #15

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1261. Min Act: -0.0001

Data Index: 552632 (C4 (Web Text))

Max Activating Token Index: 605

Text #16

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1174. Min Act: -0.0001

Data Index: 675472 (C4 (Web Text))

Max Activating Token Index: 828

Text #17

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1168. Min Act: -0.0001

Data Index: 551881 (C4 (Web Text))

Max Activating Token Index: 770

Text #18

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1188. Min Act: -0.0001

Data Index: 448114 (C4 (Web Text))

Max Activating Token Index: 308

Text #19

Max Range: 0.1477. Min Range: -0.1477

Max Act: 0.1193. Min Act: -0.0001

Data Index: 113356 (C4 (Web Text))

Max Activating Token Index: 1007