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

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

Neuron 55 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: 5.2974. Min Range: -5.2974

Max Act: 5.2974. Min Act: -0.1700

Data Index: 161770 (C4 (Web Text))

Max Activating Token Index: 712

Text #1

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.2036. Min Act: -0.1700

Data Index: 785487 (C4 (Web Text))

Max Activating Token Index: 154

Text #2

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.1411. Min Act: -0.1700

Data Index: 45492 (C4 (Web Text))

Max Activating Token Index: 256

Text #3

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.1724. Min Act: -0.1700

Data Index: 849745 (C4 (Web Text))

Max Activating Token Index: 593

Text #4

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.1411. Min Act: -0.1700

Data Index: 260449 (C4 (Web Text))

Max Activating Token Index: 694

Text #5

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.0786. Min Act: -0.1700

Data Index: 287267 (C4 (Web Text))

Max Activating Token Index: 552

Text #6

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.1099. Min Act: -0.1700

Data Index: 290343 (C4 (Web Text))

Max Activating Token Index: 921

Text #7

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.0786. Min Act: -0.1700

Data Index: 545888 (C4 (Web Text))

Max Activating Token Index: 387

Text #8

Max Range: 5.2974. Min Range: -5.2974

Max Act: 5.0161. Min Act: -0.1700

Data Index: 174765 (C4 (Web Text))

Max Activating Token Index: 268

Text #9

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8599. Min Act: -0.1700

Data Index: 933804 (C4 (Web Text))

Max Activating Token Index: 410

Text #10

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.9224. Min Act: -0.1700

Data Index: 556530 (C4 (Web Text))

Max Activating Token Index: 111

Text #11

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8286. Min Act: -0.1700

Data Index: 812638 (C4 (Web Text))

Max Activating Token Index: 290

Text #12

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8599. Min Act: -0.1700

Data Index: 1217758 (C4 (Web Text))

Max Activating Token Index: 603

Text #13

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.7661. Min Act: -0.1700

Data Index: 572775 (C4 (Web Text))

Max Activating Token Index: 945

Text #14

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8286. Min Act: -0.1700

Data Index: 615658 (C4 (Web Text))

Max Activating Token Index: 425

Text #15

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8286. Min Act: -0.1700

Data Index: 490489 (C4 (Web Text))

Max Activating Token Index: 118

Text #16

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8599. Min Act: -0.1700

Data Index: 391163 (C4 (Web Text))

Max Activating Token Index: 804

Text #17

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.8599. Min Act: -0.1700

Data Index: 1130407 (C4 (Web Text))

Max Activating Token Index: 907

Text #18

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.7661. Min Act: -0.1700

Data Index: 1002400 (C4 (Web Text))

Max Activating Token Index: 371

Text #19

Max Range: 5.2974. Min Range: -5.2974

Max Act: 4.7349. Min Act: -0.1700

Data Index: 58123 (C4 (Web Text))

Max Activating Token Index: 225