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

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

Neuron 135 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: 5.8499. Min Range: -5.8499

Max Act: 5.8499. Min Act: -0.1700

Data Index: 1433029 (Python Code)

Max Activating Token Index: 609

Text #1

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.7249. Min Act: -0.1700

Data Index: 1591042 (Python Code)

Max Activating Token Index: 329

Text #2

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.7562. Min Act: -0.1700

Data Index: 48567 (C4 (Web Text))

Max Activating Token Index: 686

Text #3

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.5687. Min Act: -0.1700

Data Index: 1414949 (Python Code)

Max Activating Token Index: 281

Text #4

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.5530. Min Act: -0.1700

Data Index: 458717 (C4 (Web Text))

Max Activating Token Index: 435

Text #5

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.6155. Min Act: -0.1700

Data Index: 1610015 (Python Code)

Max Activating Token Index: 470

Text #6

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.6312. Min Act: -0.1700

Data Index: 696553 (C4 (Web Text))

Max Activating Token Index: 622

Text #7

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.5218. Min Act: -0.1700

Data Index: 1594009 (Python Code)

Max Activating Token Index: 633

Text #8

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.5218. Min Act: -0.1700

Data Index: 1427513 (Python Code)

Max Activating Token Index: 539

Text #9

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.4749. Min Act: -0.1700

Data Index: 1531894 (Python Code)

Max Activating Token Index: 855

Text #10

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.4749. Min Act: -0.1700

Data Index: 1361460 (Python Code)

Max Activating Token Index: 248

Text #11

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.4437. Min Act: -0.1700

Data Index: 1651794 (Python Code)

Max Activating Token Index: 810

Text #12

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.4437. Min Act: -0.1700

Data Index: 1118071 (C4 (Web Text))

Max Activating Token Index: 980

Text #13

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.3187. Min Act: -0.1700

Data Index: 1592527 (Python Code)

Max Activating Token Index: 531

Text #14

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.2874. Min Act: -0.1700

Data Index: 1602363 (Python Code)

Max Activating Token Index: 826

Text #15

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.3968. Min Act: -0.1700

Data Index: 1439720 (Python Code)

Max Activating Token Index: 687

Text #16

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.2718. Min Act: -0.1700

Data Index: 1050081 (C4 (Web Text))

Max Activating Token Index: 876

Text #17

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.3187. Min Act: -0.1700

Data Index: 1419207 (Python Code)

Max Activating Token Index: 503

Text #18

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.3187. Min Act: -0.1700

Data Index: 613766 (C4 (Web Text))

Max Activating Token Index: 915

Text #19

Max Range: 5.8499. Min Range: -5.8499

Max Act: 5.3343. Min Act: -0.1700

Data Index: 16568 (C4 (Web Text))

Max Activating Token Index: 842