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

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

Neuron 1690 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: 0.3744. Min Range: -0.3744

Max Act: 0.3744. Min Act: -0.0001

Data Index: 236649 (C4 (Web Text))

Max Activating Token Index: 679

Text #1

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3447. Min Act: -0.0001

Data Index: 724973 (C4 (Web Text))

Max Activating Token Index: 91

Text #2

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3132. Min Act: -0.0001

Data Index: 1280779 (C4 (Web Text))

Max Activating Token Index: 351

Text #3

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3577. Min Act: -0.0001

Data Index: 316559 (C4 (Web Text))

Max Activating Token Index: 488

Text #4

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3279. Min Act: -0.0001

Data Index: 14601 (C4 (Web Text))

Max Activating Token Index: 727

Text #5

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2987. Min Act: -0.0001

Data Index: 443423 (C4 (Web Text))

Max Activating Token Index: 333

Text #6

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2979. Min Act: -0.0001

Data Index: 252283 (C4 (Web Text))

Max Activating Token Index: 312

Text #7

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3112. Min Act: -0.0001

Data Index: 246689 (C4 (Web Text))

Max Activating Token Index: 793

Text #8

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2921. Min Act: -0.0001

Data Index: 61088 (C4 (Web Text))

Max Activating Token Index: 293

Text #9

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3233. Min Act: -0.0001

Data Index: 1024060 (C4 (Web Text))

Max Activating Token Index: 802

Text #10

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3120. Min Act: -0.0001

Data Index: 906924 (C4 (Web Text))

Max Activating Token Index: 191

Text #11

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2744. Min Act: -0.0001

Data Index: 1458120 (Python Code)

Max Activating Token Index: 163

Text #12

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3076. Min Act: -0.0001

Data Index: 198871 (C4 (Web Text))

Max Activating Token Index: 947

Text #13

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2785. Min Act: -0.0001

Data Index: 771613 (C4 (Web Text))

Max Activating Token Index: 80

Text #14

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2699. Min Act: -0.0001

Data Index: 1516928 (Python Code)

Max Activating Token Index: 527

Text #15

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2601. Min Act: -0.0001

Data Index: 486125 (C4 (Web Text))

Max Activating Token Index: 154

Text #16

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.3047. Min Act: -0.0001

Data Index: 479983 (C4 (Web Text))

Max Activating Token Index: 583

Text #17

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2934. Min Act: -0.0001

Data Index: 1105773 (C4 (Web Text))

Max Activating Token Index: 383

Text #18

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2889. Min Act: -0.0001

Data Index: 1057361 (C4 (Web Text))

Max Activating Token Index: 678

Text #19

Max Range: 0.3744. Min Range: -0.3744

Max Act: 0.2585. Min Act: -0.0001

Data Index: 1388517 (Python Code)

Max Activating Token Index: 696