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

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

Neuron 265 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: 3.9950. Min Range: -3.9950

Max Act: 3.9950. Min Act: -0.1700

Data Index: 872515 (C4 (Web Text))

Max Activating Token Index: 1013

Text #1

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.9950. Min Act: -0.1700

Data Index: 1417249 (Python Code)

Max Activating Token Index: 559

Text #2

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.9950. Min Act: -0.1700

Data Index: 1410003 (Python Code)

Max Activating Token Index: 621

Text #3

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.9950. Min Act: -0.1700

Data Index: 754213 (C4 (Web Text))

Max Activating Token Index: 182

Text #4

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 1580578 (Python Code)

Max Activating Token Index: 586

Text #5

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 4297 (C4 (Web Text))

Max Activating Token Index: 582

Text #6

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 206314 (C4 (Web Text))

Max Activating Token Index: 78

Text #7

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8074. Min Act: -0.1700

Data Index: 1518708 (Python Code)

Max Activating Token Index: 830

Text #8

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 727092 (C4 (Web Text))

Max Activating Token Index: 90

Text #9

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.9325. Min Act: -0.1700

Data Index: 242961 (C4 (Web Text))

Max Activating Token Index: 111

Text #10

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 414909 (C4 (Web Text))

Max Activating Token Index: 557

Text #11

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8074. Min Act: -0.1700

Data Index: 1476828 (Python Code)

Max Activating Token Index: 167

Text #12

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8074. Min Act: -0.1700

Data Index: 1514969 (Python Code)

Max Activating Token Index: 993

Text #13

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 1388384 (Python Code)

Max Activating Token Index: 530

Text #14

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8074. Min Act: -0.1700

Data Index: 168542 (C4 (Web Text))

Max Activating Token Index: 784

Text #15

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.7448. Min Act: -0.1700

Data Index: 853082 (C4 (Web Text))

Max Activating Token Index: 634

Text #16

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.8699. Min Act: -0.1700

Data Index: 1026434 (C4 (Web Text))

Max Activating Token Index: 1011

Text #17

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.7448. Min Act: -0.1700

Data Index: 264044 (C4 (Web Text))

Max Activating Token Index: 114

Text #18

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.7448. Min Act: -0.1700

Data Index: 936591 (C4 (Web Text))

Max Activating Token Index: 137

Text #19

Max Range: 3.9950. Min Range: -3.9950

Max Act: 3.7448. Min Act: -0.1700

Data Index: 1474067 (Python Code)

Max Activating Token Index: 941