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

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

Neuron 3039 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: 0.7047. Min Range: -0.7047

Max Act: 0.7047. Min Act: -0.0001

Data Index: 1389101 (Python Code)

Max Activating Token Index: 859

Text #1

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5868. Min Act: -0.0001

Data Index: 410485 (C4 (Web Text))

Max Activating Token Index: 46

Text #2

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5813. Min Act: -0.0001

Data Index: 331644 (C4 (Web Text))

Max Activating Token Index: 238

Text #3

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5314. Min Act: -0.0001

Data Index: 1276658 (C4 (Web Text))

Max Activating Token Index: 475

Text #4

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5385. Min Act: -0.0001

Data Index: 1020532 (C4 (Web Text))

Max Activating Token Index: 648

Text #5

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5512. Min Act: -0.0001

Data Index: 896044 (C4 (Web Text))

Max Activating Token Index: 631

Text #6

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5310. Min Act: -0.0001

Data Index: 1504842 (Python Code)

Max Activating Token Index: 833

Text #7

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5097. Min Act: -0.0001

Data Index: 636808 (C4 (Web Text))

Max Activating Token Index: 529

Text #8

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5331. Min Act: -0.0000

Data Index: 145259 (C4 (Web Text))

Max Activating Token Index: 748

Text #9

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.5119. Min Act: -0.0001

Data Index: 656786 (C4 (Web Text))

Max Activating Token Index: 687

Text #10

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4890. Min Act: -0.0001

Data Index: 1401003 (Python Code)

Max Activating Token Index: 434

Text #11

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4965. Min Act: -0.0001

Data Index: 826006 (C4 (Web Text))

Max Activating Token Index: 771

Text #12

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4557. Min Act: -0.0001

Data Index: 916134 (C4 (Web Text))

Max Activating Token Index: 684

Text #13

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4312. Min Act: -0.0001

Data Index: 893556 (C4 (Web Text))

Max Activating Token Index: 389

Text #14

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4522. Min Act: -0.0000

Data Index: 663914 (C4 (Web Text))

Max Activating Token Index: 381

Text #15

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4722. Min Act: -0.0001

Data Index: 1452842 (Python Code)

Max Activating Token Index: 1001

Text #16

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.3747. Min Act: -0.0001

Data Index: 793634 (C4 (Web Text))

Max Activating Token Index: 309

Text #17

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4333. Min Act: -0.0001

Data Index: 631546 (C4 (Web Text))

Max Activating Token Index: 665

Text #18

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4416. Min Act: -0.0001

Data Index: 185734 (C4 (Web Text))

Max Activating Token Index: 568

Text #19

Max Range: 0.7047. Min Range: -0.7047

Max Act: 0.4493. Min Act: -0.0001

Data Index: 584569 (C4 (Web Text))

Max Activating Token Index: 528