Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 2774 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1273. Min Act: -0.0003

Data Index: 1562034 (Python Code)

Max Activating Token Index: 575

Text #1

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1377. Min Act: -0.0003

Data Index: 1139854 (C4 (Web Text))

Max Activating Token Index: 956

Text #2

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1170. Min Act: -0.0003

Data Index: 1027856 (C4 (Web Text))

Max Activating Token Index: 405

Text #3

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1426. Min Act: -0.0003

Data Index: 823563 (C4 (Web Text))

Max Activating Token Index: 149

Text #4

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1351. Min Act: -0.0003

Data Index: 717467 (C4 (Web Text))

Max Activating Token Index: 925

Text #5

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1439. Min Act: -0.0003

Data Index: 400785 (C4 (Web Text))

Max Activating Token Index: 768

Text #6

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1241. Min Act: -0.0003

Data Index: 1587107 (Python Code)

Max Activating Token Index: 44

Text #7

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1193. Min Act: -0.0003

Data Index: 1114776 (C4 (Web Text))

Max Activating Token Index: 622

Text #8

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1276. Min Act: -0.0003

Data Index: 1441608 (Python Code)

Max Activating Token Index: 246

Text #9

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1206. Min Act: -0.0003

Data Index: 1590675 (Python Code)

Max Activating Token Index: 37

Text #10

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1288. Min Act: -0.0003

Data Index: 1324761 (C4 (Web Text))

Max Activating Token Index: 708

Text #11

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1239. Min Act: -0.0003

Data Index: 657576 (C4 (Web Text))

Max Activating Token Index: 234

Text #12

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1156. Min Act: -0.0003

Data Index: 830367 (C4 (Web Text))

Max Activating Token Index: 887

Text #13

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1278. Min Act: -0.0003

Data Index: 1506503 (Python Code)

Max Activating Token Index: 44

Text #14

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1095. Min Act: -0.0003

Data Index: 351851 (C4 (Web Text))

Max Activating Token Index: 113

Text #15

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1283. Min Act: -0.0003

Data Index: 1353899 (C4 (Web Text))

Max Activating Token Index: 721

Text #16

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1259. Min Act: -0.0003

Data Index: 1448572 (Python Code)

Max Activating Token Index: 540

Text #17

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1212. Min Act: -0.0003

Data Index: 137263 (C4 (Web Text))

Max Activating Token Index: 92

Text #18

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1154. Min Act: -0.0003

Data Index: 267506 (C4 (Web Text))

Max Activating Token Index: 528

Text #19

Max Range: 0.1439. Min Range: -0.1439

Max Act: 0.1164. Min Act: -0.0003

Data Index: 52436 (C4 (Web Text))

Max Activating Token Index: 391