Model: GELU Model: 4 Layers, 2048 Neurons per Layer

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

Neuron 1883 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-4l')

Text #0

Max Range: 4.1993. Min Range: -4.1993

Max Act: 4.1681. Min Act: -0.1700

Data Index: 833926 (C4 (Web Text))

Max Activating Token Index: 953

Text #1

Max Range: 4.1993. Min Range: -4.1993

Max Act: 4.1993. Min Act: -0.1700

Data Index: 1008951 (C4 (Web Text))

Max Activating Token Index: 725

Text #2

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9805. Min Act: -0.1700

Data Index: 362835 (C4 (Web Text))

Max Activating Token Index: 430

Text #3

Max Range: 4.1993. Min Range: -4.1993

Max Act: 4.0430. Min Act: -0.1700

Data Index: 1027938 (C4 (Web Text))

Max Activating Token Index: 297

Text #4

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9805. Min Act: -0.1700

Data Index: 944776 (C4 (Web Text))

Max Activating Token Index: 889

Text #5

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9492. Min Act: -0.1700

Data Index: 1114563 (C4 (Web Text))

Max Activating Token Index: 818

Text #6

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9492. Min Act: -0.1700

Data Index: 797921 (C4 (Web Text))

Max Activating Token Index: 683

Text #7

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9492. Min Act: -0.1700

Data Index: 906802 (C4 (Web Text))

Max Activating Token Index: 982

Text #8

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.9179. Min Act: -0.1700

Data Index: 705863 (C4 (Web Text))

Max Activating Token Index: 182

Text #9

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.8867. Min Act: -0.1700

Data Index: 480838 (C4 (Web Text))

Max Activating Token Index: 998

Text #10

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.7928. Min Act: -0.1700

Data Index: 766900 (C4 (Web Text))

Max Activating Token Index: 802

Text #11

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.8867. Min Act: -0.1700

Data Index: 335471 (C4 (Web Text))

Max Activating Token Index: 554

Text #12

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.8241. Min Act: -0.1700

Data Index: 668714 (C4 (Web Text))

Max Activating Token Index: 897

Text #13

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.8554. Min Act: -0.1700

Data Index: 619830 (C4 (Web Text))

Max Activating Token Index: 864

Text #14

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.8085. Min Act: -0.1700

Data Index: 234600 (C4 (Web Text))

Max Activating Token Index: 990

Text #15

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.7928. Min Act: -0.1700

Data Index: 831063 (C4 (Web Text))

Max Activating Token Index: 502

Text #16

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.7928. Min Act: -0.1700

Data Index: 1039614 (C4 (Web Text))

Max Activating Token Index: 693

Text #17

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.6990. Min Act: -0.1700

Data Index: 945113 (C4 (Web Text))

Max Activating Token Index: 257

Text #18

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.7303. Min Act: -0.1700

Data Index: 862645 (C4 (Web Text))

Max Activating Token Index: 420

Text #19

Max Range: 4.1993. Min Range: -4.1993

Max Act: 3.7459. Min Act: -0.1700

Data Index: 942171 (C4 (Web Text))

Max Activating Token Index: 111