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

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

Neuron 2755 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3925. Min Act: -0.0001

Data Index: 504595 (C4 (Web Text))

Max Activating Token Index: 380

Text #1

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3702. Min Act: -0.0001

Data Index: 1515465 (Python Code)

Max Activating Token Index: 835

Text #2

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3984. Min Act: -0.0000

Data Index: 180927 (C4 (Web Text))

Max Activating Token Index: 90

Text #3

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3692. Min Act: -0.0001

Data Index: 788560 (C4 (Web Text))

Max Activating Token Index: 933

Text #4

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3503. Min Act: -0.0000

Data Index: 43875 (C4 (Web Text))

Max Activating Token Index: 103

Text #5

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3346. Min Act: -0.0001

Data Index: 1148835 (C4 (Web Text))

Max Activating Token Index: 853

Text #6

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3485. Min Act: -0.0001

Data Index: 892853 (C4 (Web Text))

Max Activating Token Index: 88

Text #7

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3643. Min Act: -0.0001

Data Index: 1299161 (C4 (Web Text))

Max Activating Token Index: 349

Text #8

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3425. Min Act: -0.0001

Data Index: 641150 (C4 (Web Text))

Max Activating Token Index: 921

Text #9

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3376. Min Act: -0.0000

Data Index: 664330 (C4 (Web Text))

Max Activating Token Index: 57

Text #10

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3246. Min Act: -0.0001

Data Index: 1613500 (Python Code)

Max Activating Token Index: 512

Text #11

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3303. Min Act: -0.0001

Data Index: 1202704 (C4 (Web Text))

Max Activating Token Index: 390

Text #12

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3508. Min Act: -0.0001

Data Index: 1530753 (Python Code)

Max Activating Token Index: 649

Text #13

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3148. Min Act: -0.0001

Data Index: 612828 (C4 (Web Text))

Max Activating Token Index: 17

Text #14

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3328. Min Act: -0.0001

Data Index: 355995 (C4 (Web Text))

Max Activating Token Index: 1019

Text #15

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3129. Min Act: -0.0001

Data Index: 778497 (C4 (Web Text))

Max Activating Token Index: 615

Text #16

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3179. Min Act: -0.0001

Data Index: 1176516 (C4 (Web Text))

Max Activating Token Index: 664

Text #17

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3274. Min Act: -0.0001

Data Index: 598163 (C4 (Web Text))

Max Activating Token Index: 1017

Text #18

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3148. Min Act: -0.0000

Data Index: 867781 (C4 (Web Text))

Max Activating Token Index: 505

Text #19

Max Range: 0.3984. Min Range: -0.3984

Max Act: 0.3116. Min Act: -0.0000

Data Index: 382599 (C4 (Web Text))

Max Activating Token Index: 933