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

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

Neuron 342 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: 8.8060. Min Range: -8.8060

Max Act: 8.8060. Min Act: -0.1700

Data Index: 189113 (C4 (Web Text))

Max Activating Token Index: 420

Text #1

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.7435. Min Act: -0.1700

Data Index: 267923 (C4 (Web Text))

Max Activating Token Index: 261

Text #2

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.6498. Min Act: -0.1700

Data Index: 1492313 (Python Code)

Max Activating Token Index: 347

Text #3

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.6185. Min Act: -0.1700

Data Index: 424538 (C4 (Web Text))

Max Activating Token Index: 460

Text #4

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.2748. Min Act: -0.1700

Data Index: 1165865 (C4 (Web Text))

Max Activating Token Index: 949

Text #5

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.1498. Min Act: -0.1700

Data Index: 1305053 (C4 (Web Text))

Max Activating Token Index: 266

Text #6

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.1185. Min Act: -0.1700

Data Index: 1462856 (Python Code)

Max Activating Token Index: 479

Text #7

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.0560. Min Act: -0.1700

Data Index: 1590818 (Python Code)

Max Activating Token Index: 896

Text #8

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.9935. Min Act: -0.1700

Data Index: 1526449 (Python Code)

Max Activating Token Index: 443

Text #9

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.0873. Min Act: -0.1700

Data Index: 1549686 (Python Code)

Max Activating Token Index: 291

Text #10

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.9935. Min Act: -0.1700

Data Index: 341881 (C4 (Web Text))

Max Activating Token Index: 387

Text #11

Max Range: 8.8060. Min Range: -8.8060

Max Act: 8.0248. Min Act: -0.1700

Data Index: 1380221 (Python Code)

Max Activating Token Index: 1022

Text #12

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.9623. Min Act: -0.1700

Data Index: 1594381 (Python Code)

Max Activating Token Index: 210

Text #13

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.8373. Min Act: -0.1700

Data Index: 1492701 (Python Code)

Max Activating Token Index: 822

Text #14

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.9623. Min Act: -0.1700

Data Index: 1399188 (Python Code)

Max Activating Token Index: 315

Text #15

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.8998. Min Act: -0.1700

Data Index: 1627226 (Python Code)

Max Activating Token Index: 717

Text #16

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.8685. Min Act: -0.1700

Data Index: 435081 (C4 (Web Text))

Max Activating Token Index: 268

Text #17

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.8060. Min Act: -0.1700

Data Index: 1452038 (Python Code)

Max Activating Token Index: 285

Text #18

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.8060. Min Act: -0.1700

Data Index: 1410541 (Python Code)

Max Activating Token Index: 602

Text #19

Max Range: 8.8060. Min Range: -8.8060

Max Act: 7.7748. Min Act: -0.1700

Data Index: 1216810 (C4 (Web Text))

Max Activating Token Index: 464