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

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

Neuron 2044 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: 5.2218. Min Range: -5.2218

Max Act: 5.2218. Min Act: -0.1700

Data Index: 1116539 (C4 (Web Text))

Max Activating Token Index: 708

Text #1

Max Range: 5.2218. Min Range: -5.2218

Max Act: 5.1280. Min Act: -0.1700

Data Index: 1476512 (Python Code)

Max Activating Token Index: 769

Text #2

Max Range: 5.2218. Min Range: -5.2218

Max Act: 5.1280. Min Act: -0.1700

Data Index: 984367 (C4 (Web Text))

Max Activating Token Index: 182

Text #3

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.9718. Min Act: -0.1700

Data Index: 1510282 (Python Code)

Max Activating Token Index: 966

Text #4

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.9405. Min Act: -0.1700

Data Index: 1583140 (Python Code)

Max Activating Token Index: 728

Text #5

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8780. Min Act: -0.1700

Data Index: 619488 (C4 (Web Text))

Max Activating Token Index: 347

Text #6

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.9093. Min Act: -0.1700

Data Index: 1194582 (C4 (Web Text))

Max Activating Token Index: 309

Text #7

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8780. Min Act: -0.1700

Data Index: 1380447 (Python Code)

Max Activating Token Index: 943

Text #8

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8780. Min Act: -0.1700

Data Index: 1432244 (Python Code)

Max Activating Token Index: 902

Text #9

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1413539 (Python Code)

Max Activating Token Index: 722

Text #10

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8468. Min Act: -0.1700

Data Index: 1373363 (Python Code)

Max Activating Token Index: 437

Text #11

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8780. Min Act: -0.1700

Data Index: 1443459 (Python Code)

Max Activating Token Index: 608

Text #12

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1350762 (C4 (Web Text))

Max Activating Token Index: 378

Text #13

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8468. Min Act: -0.1700

Data Index: 1568946 (Python Code)

Max Activating Token Index: 840

Text #14

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1406861 (Python Code)

Max Activating Token Index: 646

Text #15

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8468. Min Act: -0.1700

Data Index: 1451015 (Python Code)

Max Activating Token Index: 477

Text #16

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1465249 (Python Code)

Max Activating Token Index: 646

Text #17

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1437879 (Python Code)

Max Activating Token Index: 239

Text #18

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8468. Min Act: -0.1700

Data Index: 1360173 (Python Code)

Max Activating Token Index: 886

Text #19

Max Range: 5.2218. Min Range: -5.2218

Max Act: 4.8155. Min Act: -0.1700

Data Index: 1628792 (Python Code)

Max Activating Token Index: 467