Model: GELU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 1064 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('gelu-3l')

Text #0

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.8585. Min Act: -0.1700

Data Index: 1438105 (Python Code)

Max Activating Token Index: 410

Text #1

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.8585. Min Act: -0.1700

Data Index: 502849 (C4 (Web Text))

Max Activating Token Index: 420

Text #2

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.7803. Min Act: -0.1700

Data Index: 1576990 (Python Code)

Max Activating Token Index: 883

Text #3

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6238. Min Act: -0.1700

Data Index: 442620 (C4 (Web Text))

Max Activating Token Index: 501

Text #4

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1615073 (Python Code)

Max Activating Token Index: 741

Text #5

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6551. Min Act: -0.1700

Data Index: 1642972 (Python Code)

Max Activating Token Index: 352

Text #6

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.5925. Min Act: -0.1700

Data Index: 1454061 (Python Code)

Max Activating Token Index: 904

Text #7

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6864. Min Act: -0.1700

Data Index: 1389904 (Python Code)

Max Activating Token Index: 403

Text #8

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6551. Min Act: -0.1700

Data Index: 1590642 (Python Code)

Max Activating Token Index: 959

Text #9

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6551. Min Act: -0.1700

Data Index: 436791 (C4 (Web Text))

Max Activating Token Index: 259

Text #10

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6238. Min Act: -0.1700

Data Index: 682339 (C4 (Web Text))

Max Activating Token Index: 584

Text #11

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.7021. Min Act: -0.1700

Data Index: 1554731 (Python Code)

Max Activating Token Index: 337

Text #12

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.5925. Min Act: -0.1700

Data Index: 1129705 (C4 (Web Text))

Max Activating Token Index: 447

Text #13

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6864. Min Act: -0.1700

Data Index: 769991 (C4 (Web Text))

Max Activating Token Index: 898

Text #14

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6864. Min Act: -0.1700

Data Index: 1436279 (Python Code)

Max Activating Token Index: 156

Text #15

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6238. Min Act: -0.1700

Data Index: 1437207 (Python Code)

Max Activating Token Index: 325

Text #16

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1403948 (Python Code)

Max Activating Token Index: 938

Text #17

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1402349 (Python Code)

Max Activating Token Index: 105

Text #18

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6395. Min Act: -0.1700

Data Index: 1365275 (Python Code)

Max Activating Token Index: 300

Text #19

Max Range: 3.8585. Min Range: -3.8585

Max Act: 3.6238. Min Act: -0.1700

Data Index: 1362583 (Python Code)

Max Activating Token Index: 521