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

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

Neuron 312 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.8492. Min Act: -0.1700

Data Index: 362618 (C4 (Web Text))

Max Activating Token Index: 320

Text #1

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.8492. Min Act: -0.1700

Data Index: 680264 (C4 (Web Text))

Max Activating Token Index: 643

Text #2

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.7554. Min Act: -0.1700

Data Index: 324101 (C4 (Web Text))

Max Activating Token Index: 367

Text #3

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.7242. Min Act: -0.1700

Data Index: 278769 (C4 (Web Text))

Max Activating Token Index: 223

Text #4

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6929. Min Act: -0.1700

Data Index: 598526 (C4 (Web Text))

Max Activating Token Index: 162

Text #5

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6304. Min Act: -0.1700

Data Index: 1538065 (Python Code)

Max Activating Token Index: 21

Text #6

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6617. Min Act: -0.1700

Data Index: 865765 (C4 (Web Text))

Max Activating Token Index: 65

Text #7

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5992. Min Act: -0.1700

Data Index: 809334 (C4 (Web Text))

Max Activating Token Index: 946

Text #8

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6617. Min Act: -0.1700

Data Index: 767144 (C4 (Web Text))

Max Activating Token Index: 609

Text #9

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6304. Min Act: -0.1700

Data Index: 1626306 (Python Code)

Max Activating Token Index: 23

Text #10

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.6617. Min Act: -0.1700

Data Index: 1208205 (C4 (Web Text))

Max Activating Token Index: 161

Text #11

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5367. Min Act: -0.1700

Data Index: 73629 (C4 (Web Text))

Max Activating Token Index: 274

Text #12

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.4429. Min Act: -0.1700

Data Index: 1175919 (C4 (Web Text))

Max Activating Token Index: 441

Text #13

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5679. Min Act: -0.1700

Data Index: 853110 (C4 (Web Text))

Max Activating Token Index: 270

Text #14

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5367. Min Act: -0.1700

Data Index: 1065300 (C4 (Web Text))

Max Activating Token Index: 165

Text #15

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5367. Min Act: -0.1700

Data Index: 542248 (C4 (Web Text))

Max Activating Token Index: 993

Text #16

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5679. Min Act: -0.1700

Data Index: 1624395 (Python Code)

Max Activating Token Index: 34

Text #17

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5367. Min Act: -0.1700

Data Index: 1611431 (Python Code)

Max Activating Token Index: 22

Text #18

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5054. Min Act: -0.1700

Data Index: 1313072 (C4 (Web Text))

Max Activating Token Index: 436

Text #19

Max Range: 5.8492. Min Range: -5.8492

Max Act: 5.5679. Min Act: -0.1700

Data Index: 1322705 (C4 (Web Text))

Max Activating Token Index: 189