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

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

Neuron 2845 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.4489. Min Act: -0.0001

Data Index: 273491 (C4 (Web Text))

Max Activating Token Index: 157

Text #1

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.4510. Min Act: -0.0001

Data Index: 353432 (C4 (Web Text))

Max Activating Token Index: 531

Text #2

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.4082. Min Act: -0.0001

Data Index: 287160 (C4 (Web Text))

Max Activating Token Index: 71

Text #3

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.3688. Min Act: -0.0001

Data Index: 1344974 (C4 (Web Text))

Max Activating Token Index: 458

Text #4

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.1626. Min Act: -0.0002

Data Index: 49428 (C4 (Web Text))

Max Activating Token Index: 991

Text #5

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.1647. Min Act: -0.0002

Data Index: 931287 (C4 (Web Text))

Max Activating Token Index: 175

Text #6

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.1071. Min Act: -0.0001

Data Index: 245570 (C4 (Web Text))

Max Activating Token Index: 122

Text #7

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.0464. Min Act: -0.0001

Data Index: 747914 (C4 (Web Text))

Max Activating Token Index: 952

Text #8

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.0274. Min Act: -0.0001

Data Index: 721386 (C4 (Web Text))

Max Activating Token Index: 397

Text #9

Max Range: 1.4510. Min Range: -1.4510

Max Act: 1.0584. Min Act: -0.0002

Data Index: 941350 (C4 (Web Text))

Max Activating Token Index: 126

Text #10

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.9684. Min Act: -0.0001

Data Index: 1202284 (C4 (Web Text))

Max Activating Token Index: 617

Text #11

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8561. Min Act: -0.0001

Data Index: 378636 (C4 (Web Text))

Max Activating Token Index: 214

Text #12

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.9378. Min Act: -0.0001

Data Index: 232569 (C4 (Web Text))

Max Activating Token Index: 976

Text #13

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.9203. Min Act: -0.0001

Data Index: 108126 (C4 (Web Text))

Max Activating Token Index: 835

Text #14

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8520. Min Act: -0.0001

Data Index: 50396 (C4 (Web Text))

Max Activating Token Index: 591

Text #15

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8603. Min Act: -0.0001

Data Index: 1388296 (Python Code)

Max Activating Token Index: 624

Text #16

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8831. Min Act: -0.0001

Data Index: 571485 (C4 (Web Text))

Max Activating Token Index: 653

Text #17

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8415. Min Act: -0.0001

Data Index: 422529 (C4 (Web Text))

Max Activating Token Index: 79

Text #18

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.8265. Min Act: -0.0001

Data Index: 333037 (C4 (Web Text))

Max Activating Token Index: 265

Text #19

Max Range: 1.4510. Min Range: -1.4510

Max Act: 0.9625. Min Act: -0.0001

Data Index: 838045 (C4 (Web Text))

Max Activating Token Index: 145