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

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

Neuron 2025 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: 3.2947. Min Range: -3.2947

Max Act: 3.2947. Min Act: -0.1700

Data Index: 39555 (C4 (Web Text))

Max Activating Token Index: 599

Text #1

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.2790. Min Act: -0.1700

Data Index: 1245579 (C4 (Web Text))

Max Activating Token Index: 756

Text #2

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.2318. Min Act: -0.1700

Data Index: 463864 (C4 (Web Text))

Max Activating Token Index: 553

Text #3

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.2475. Min Act: -0.1700

Data Index: 823770 (C4 (Web Text))

Max Activating Token Index: 448

Text #4

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.2161. Min Act: -0.1700

Data Index: 871878 (C4 (Web Text))

Max Activating Token Index: 712

Text #5

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.2161. Min Act: -0.1700

Data Index: 28179 (C4 (Web Text))

Max Activating Token Index: 506

Text #6

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1531. Min Act: -0.1700

Data Index: 1014144 (C4 (Web Text))

Max Activating Token Index: 494

Text #7

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1531. Min Act: -0.1700

Data Index: 1020940 (C4 (Web Text))

Max Activating Token Index: 247

Text #8

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1216. Min Act: -0.1700

Data Index: 1121303 (C4 (Web Text))

Max Activating Token Index: 181

Text #9

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.0901. Min Act: -0.1700

Data Index: 157635 (C4 (Web Text))

Max Activating Token Index: 673

Text #10

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1059. Min Act: -0.1700

Data Index: 268733 (C4 (Web Text))

Max Activating Token Index: 816

Text #11

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1059. Min Act: -0.1700

Data Index: 79415 (C4 (Web Text))

Max Activating Token Index: 589

Text #12

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1216. Min Act: -0.1700

Data Index: 1121056 (C4 (Web Text))

Max Activating Token Index: 357

Text #13

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1059. Min Act: -0.1700

Data Index: 427669 (C4 (Web Text))

Max Activating Token Index: 390

Text #14

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1059. Min Act: -0.1700

Data Index: 482724 (C4 (Web Text))

Max Activating Token Index: 472

Text #15

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.0901. Min Act: -0.1700

Data Index: 1216380 (C4 (Web Text))

Max Activating Token Index: 808

Text #16

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.1059. Min Act: -0.1700

Data Index: 1230037 (C4 (Web Text))

Max Activating Token Index: 649

Text #17

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.0901. Min Act: -0.1700

Data Index: 1013437 (C4 (Web Text))

Max Activating Token Index: 700

Text #18

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.0901. Min Act: -0.1700

Data Index: 614273 (C4 (Web Text))

Max Activating Token Index: 757

Text #19

Max Range: 3.2947. Min Range: -3.2947

Max Act: 3.0901. Min Act: -0.1700

Data Index: 960732 (C4 (Web Text))

Max Activating Token Index: 118