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

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

Neuron 1712 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: 8.0456. Min Range: -8.0456

Max Act: 8.0456. Min Act: -0.1700

Data Index: 511827 (C4 (Web Text))

Max Activating Token Index: 114

Text #1

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.9206. Min Act: -0.1700

Data Index: 911878 (C4 (Web Text))

Max Activating Token Index: 904

Text #2

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.9206. Min Act: -0.1700

Data Index: 1141807 (C4 (Web Text))

Max Activating Token Index: 366

Text #3

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.9206. Min Act: -0.1700

Data Index: 1190516 (C4 (Web Text))

Max Activating Token Index: 634

Text #4

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.7956. Min Act: -0.1700

Data Index: 1036908 (C4 (Web Text))

Max Activating Token Index: 129

Text #5

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.8268. Min Act: -0.1700

Data Index: 1299377 (C4 (Web Text))

Max Activating Token Index: 428

Text #6

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.8893. Min Act: -0.1700

Data Index: 899728 (C4 (Web Text))

Max Activating Token Index: 107

Text #7

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.8268. Min Act: -0.1700

Data Index: 1103358 (C4 (Web Text))

Max Activating Token Index: 508

Text #8

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.7331. Min Act: -0.1700

Data Index: 438796 (C4 (Web Text))

Max Activating Token Index: 58

Text #9

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.7799. Min Act: -0.1700

Data Index: 1340319 (C4 (Web Text))

Max Activating Token Index: 154

Text #10

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.8268. Min Act: -0.1700

Data Index: 20973 (C4 (Web Text))

Max Activating Token Index: 500

Text #11

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.6862. Min Act: -0.1692

Data Index: 813764 (C4 (Web Text))

Max Activating Token Index: 137

Text #12

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.6081. Min Act: -0.1700

Data Index: 747762 (C4 (Web Text))

Max Activating Token Index: 508

Text #13

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.6393. Min Act: -0.1692

Data Index: 582051 (C4 (Web Text))

Max Activating Token Index: 405

Text #14

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.6706. Min Act: -0.1700

Data Index: 1481213 (Python Code)

Max Activating Token Index: 834

Text #15

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.6706. Min Act: -0.1698

Data Index: 794444 (C4 (Web Text))

Max Activating Token Index: 99

Text #16

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.4987. Min Act: -0.1700

Data Index: 101086 (C4 (Web Text))

Max Activating Token Index: 241

Text #17

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.5768. Min Act: -0.1698

Data Index: 399446 (C4 (Web Text))

Max Activating Token Index: 671

Text #18

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.5299. Min Act: -0.1700

Data Index: 682771 (C4 (Web Text))

Max Activating Token Index: 263

Text #19

Max Range: 8.0456. Min Range: -8.0456

Max Act: 7.4518. Min Act: -0.1700

Data Index: 1563770 (Python Code)

Max Activating Token Index: 899