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

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

Neuron 2338 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7809. Min Act: -0.0001

Data Index: 880167 (C4 (Web Text))

Max Activating Token Index: 951

Text #1

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.8013. Min Act: -0.0000

Data Index: 80634 (C4 (Web Text))

Max Activating Token Index: 79

Text #2

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7633. Min Act: -0.0001

Data Index: 953801 (C4 (Web Text))

Max Activating Token Index: 564

Text #3

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.8311. Min Act: -0.0000

Data Index: 660683 (C4 (Web Text))

Max Activating Token Index: 569

Text #4

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7985. Min Act: -0.0000

Data Index: 1449710 (Python Code)

Max Activating Token Index: 189

Text #5

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7431. Min Act: -0.0000

Data Index: 544548 (C4 (Web Text))

Max Activating Token Index: 74

Text #6

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7550. Min Act: -0.0001

Data Index: 642033 (C4 (Web Text))

Max Activating Token Index: 910

Text #7

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6920. Min Act: -0.0001

Data Index: 658507 (C4 (Web Text))

Max Activating Token Index: 996

Text #8

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6533. Min Act: -0.0000

Data Index: 452424 (C4 (Web Text))

Max Activating Token Index: 543

Text #9

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6758. Min Act: -0.0000

Data Index: 1130793 (C4 (Web Text))

Max Activating Token Index: 916

Text #10

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7128. Min Act: -0.0000

Data Index: 1526040 (Python Code)

Max Activating Token Index: 291

Text #11

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.7319. Min Act: -0.0001

Data Index: 1032965 (C4 (Web Text))

Max Activating Token Index: 208

Text #12

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6562. Min Act: -0.0000

Data Index: 1480071 (Python Code)

Max Activating Token Index: 388

Text #13

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6592. Min Act: -0.0001

Data Index: 387533 (C4 (Web Text))

Max Activating Token Index: 199

Text #14

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.5765. Min Act: -0.0000

Data Index: 559697 (C4 (Web Text))

Max Activating Token Index: 807

Text #15

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6169. Min Act: -0.0000

Data Index: 842163 (C4 (Web Text))

Max Activating Token Index: 930

Text #16

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6488. Min Act: -0.0001

Data Index: 1247250 (C4 (Web Text))

Max Activating Token Index: 757

Text #17

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6210. Min Act: -0.0000

Data Index: 725436 (C4 (Web Text))

Max Activating Token Index: 579

Text #18

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6214. Min Act: -0.0000

Data Index: 436008 (C4 (Web Text))

Max Activating Token Index: 80

Text #19

Max Range: 0.8311. Min Range: -0.8311

Max Act: 0.6307. Min Act: -0.0000

Data Index: 363404 (C4 (Web Text))

Max Activating Token Index: 486