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

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

Neuron 670 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: 8.5131. Min Range: -8.5131

Max Act: 8.5131. Min Act: -0.1700

Data Index: 526180 (C4 (Web Text))

Max Activating Token Index: 880

Text #1

Max Range: 8.5131. Min Range: -8.5131

Max Act: 8.0756. Min Act: -0.1700

Data Index: 1143944 (C4 (Web Text))

Max Activating Token Index: 882

Text #2

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.7944. Min Act: -0.1700

Data Index: 259798 (C4 (Web Text))

Max Activating Token Index: 951

Text #3

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.7006. Min Act: -0.1700

Data Index: 623811 (C4 (Web Text))

Max Activating Token Index: 487

Text #4

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.6069. Min Act: -0.1700

Data Index: 638728 (C4 (Web Text))

Max Activating Token Index: 748

Text #5

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.5444. Min Act: -0.1700

Data Index: 364147 (C4 (Web Text))

Max Activating Token Index: 895

Text #6

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.4506. Min Act: -0.1700

Data Index: 1008987 (C4 (Web Text))

Max Activating Token Index: 188

Text #7

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.4819. Min Act: -0.1700

Data Index: 909584 (C4 (Web Text))

Max Activating Token Index: 838

Text #8

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.4194. Min Act: -0.1700

Data Index: 1140428 (C4 (Web Text))

Max Activating Token Index: 570

Text #9

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3881. Min Act: -0.1700

Data Index: 1096598 (C4 (Web Text))

Max Activating Token Index: 774

Text #10

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3569. Min Act: -0.1699

Data Index: 725348 (C4 (Web Text))

Max Activating Token Index: 912

Text #11

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3256. Min Act: -0.1700

Data Index: 1341280 (C4 (Web Text))

Max Activating Token Index: 704

Text #12

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3881. Min Act: -0.1700

Data Index: 1409741 (Python Code)

Max Activating Token Index: 589

Text #13

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3569. Min Act: -0.1700

Data Index: 1224866 (C4 (Web Text))

Max Activating Token Index: 179

Text #14

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3569. Min Act: -0.1700

Data Index: 502039 (C4 (Web Text))

Max Activating Token Index: 657

Text #15

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3881. Min Act: -0.1700

Data Index: 805434 (C4 (Web Text))

Max Activating Token Index: 1014

Text #16

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3569. Min Act: -0.1700

Data Index: 405844 (C4 (Web Text))

Max Activating Token Index: 728

Text #17

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3256. Min Act: -0.1700

Data Index: 765551 (C4 (Web Text))

Max Activating Token Index: 230

Text #18

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.3256. Min Act: -0.1700

Data Index: 49517 (C4 (Web Text))

Max Activating Token Index: 882

Text #19

Max Range: 8.5131. Min Range: -8.5131

Max Act: 7.2319. Min Act: -0.1700

Data Index: 991201 (C4 (Web Text))

Max Activating Token Index: 830