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

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

Neuron 3200 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.5777. Min Act: -0.0001

Data Index: 23259 (C4 (Web Text))

Max Activating Token Index: 587

Text #1

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.5409. Min Act: -0.0001

Data Index: 486724 (C4 (Web Text))

Max Activating Token Index: 52

Text #2

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.5110. Min Act: -0.0000

Data Index: 459451 (C4 (Web Text))

Max Activating Token Index: 546

Text #3

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.5026. Min Act: -0.0000

Data Index: 475674 (C4 (Web Text))

Max Activating Token Index: 377

Text #4

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4787. Min Act: -0.0000

Data Index: 321823 (C4 (Web Text))

Max Activating Token Index: 876

Text #5

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4660. Min Act: -0.0000

Data Index: 198878 (C4 (Web Text))

Max Activating Token Index: 666

Text #6

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4656. Min Act: -0.0000

Data Index: 591713 (C4 (Web Text))

Max Activating Token Index: 224

Text #7

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4783. Min Act: -0.0000

Data Index: 93511 (C4 (Web Text))

Max Activating Token Index: 815

Text #8

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4675. Min Act: -0.0001

Data Index: 298734 (C4 (Web Text))

Max Activating Token Index: 757

Text #9

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4700. Min Act: -0.0000

Data Index: 888004 (C4 (Web Text))

Max Activating Token Index: 685

Text #10

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4694. Min Act: -0.0001

Data Index: 280174 (C4 (Web Text))

Max Activating Token Index: 127

Text #11

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4244. Min Act: -0.0000

Data Index: 575474 (C4 (Web Text))

Max Activating Token Index: 842

Text #12

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4525. Min Act: -0.0000

Data Index: 709649 (C4 (Web Text))

Max Activating Token Index: 879

Text #13

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4715. Min Act: -0.0000

Data Index: 86642 (C4 (Web Text))

Max Activating Token Index: 840

Text #14

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4308. Min Act: -0.0000

Data Index: 152022 (C4 (Web Text))

Max Activating Token Index: 116

Text #15

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4430. Min Act: -0.0000

Data Index: 757271 (C4 (Web Text))

Max Activating Token Index: 239

Text #16

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4296. Min Act: -0.0000

Data Index: 443969 (C4 (Web Text))

Max Activating Token Index: 536

Text #17

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4593. Min Act: -0.0000

Data Index: 176454 (C4 (Web Text))

Max Activating Token Index: 893

Text #18

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4698. Min Act: -0.0000

Data Index: 563668 (C4 (Web Text))

Max Activating Token Index: 864

Text #19

Max Range: 0.5777. Min Range: -0.5777

Max Act: 0.4385. Min Act: -0.0001

Data Index: 1031468 (C4 (Web Text))

Max Activating Token Index: 403