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

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

Neuron 2026 in Layer 1

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.5231. Min Act: -0.1700

Data Index: 1325760 (C4 (Web Text))

Max Activating Token Index: 857

Text #1

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.3506. Min Act: -0.1700

Data Index: 328752 (C4 (Web Text))

Max Activating Token Index: 381

Text #2

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.3349. Min Act: -0.1700

Data Index: 1026683 (C4 (Web Text))

Max Activating Token Index: 286

Text #3

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2878. Min Act: -0.1700

Data Index: 1018650 (C4 (Web Text))

Max Activating Token Index: 409

Text #4

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2407. Min Act: -0.1700

Data Index: 200616 (C4 (Web Text))

Max Activating Token Index: 805

Text #5

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2564. Min Act: -0.1700

Data Index: 647030 (C4 (Web Text))

Max Activating Token Index: 330

Text #6

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2564. Min Act: -0.1700

Data Index: 260320 (C4 (Web Text))

Max Activating Token Index: 365

Text #7

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2564. Min Act: -0.1700

Data Index: 234954 (C4 (Web Text))

Max Activating Token Index: 686

Text #8

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2721. Min Act: -0.1700

Data Index: 1046156 (C4 (Web Text))

Max Activating Token Index: 470

Text #9

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2564. Min Act: -0.1700

Data Index: 859257 (C4 (Web Text))

Max Activating Token Index: 410

Text #10

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2564. Min Act: -0.1700

Data Index: 282213 (C4 (Web Text))

Max Activating Token Index: 881

Text #11

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2250. Min Act: -0.1700

Data Index: 507470 (C4 (Web Text))

Max Activating Token Index: 711

Text #12

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2250. Min Act: -0.1700

Data Index: 356312 (C4 (Web Text))

Max Activating Token Index: 976

Text #13

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2092. Min Act: -0.1700

Data Index: 196887 (C4 (Web Text))

Max Activating Token Index: 634

Text #14

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2250. Min Act: -0.1700

Data Index: 896356 (C4 (Web Text))

Max Activating Token Index: 884

Text #15

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2092. Min Act: -0.1700

Data Index: 1143452 (C4 (Web Text))

Max Activating Token Index: 346

Text #16

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.1935. Min Act: -0.1700

Data Index: 951232 (C4 (Web Text))

Max Activating Token Index: 79

Text #17

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.2250. Min Act: -0.1700

Data Index: 561993 (C4 (Web Text))

Max Activating Token Index: 790

Text #18

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.1935. Min Act: -0.1700

Data Index: 160634 (C4 (Web Text))

Max Activating Token Index: 295

Text #19

Max Range: 3.5231. Min Range: -3.5231

Max Act: 3.1778. Min Act: -0.1700

Data Index: 1016855 (C4 (Web Text))

Max Activating Token Index: 976