Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 6064 in Layer 10

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.5672. Min Act: -0.0000

Data Index: 1637672 (Python Code)

Max Activating Token Index: 290

Text #1

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.5392. Min Act: -0.0000

Data Index: 338151 (C4 (Web Text))

Max Activating Token Index: 474

Text #2

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4790. Min Act: -0.0000

Data Index: 540969 (C4 (Web Text))

Max Activating Token Index: 647

Text #3

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.5010. Min Act: -0.0000

Data Index: 1548578 (Python Code)

Max Activating Token Index: 272

Text #4

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4634. Min Act: -0.0000

Data Index: 1392417 (Python Code)

Max Activating Token Index: 258

Text #5

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4796. Min Act: -0.0000

Data Index: 581021 (C4 (Web Text))

Max Activating Token Index: 701

Text #6

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4868. Min Act: -0.0000

Data Index: 1527568 (Python Code)

Max Activating Token Index: 424

Text #7

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4440. Min Act: -0.0000

Data Index: 202241 (C4 (Web Text))

Max Activating Token Index: 691

Text #8

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4375. Min Act: -0.0000

Data Index: 364849 (C4 (Web Text))

Max Activating Token Index: 84

Text #9

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4208. Min Act: -0.0000

Data Index: 1460925 (Python Code)

Max Activating Token Index: 963

Text #10

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4270. Min Act: -0.0000

Data Index: 640914 (C4 (Web Text))

Max Activating Token Index: 213

Text #11

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4154. Min Act: -0.0000

Data Index: 1509202 (Python Code)

Max Activating Token Index: 259

Text #12

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4037. Min Act: -0.0000

Data Index: 234248 (C4 (Web Text))

Max Activating Token Index: 381

Text #13

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4001. Min Act: -0.0000

Data Index: 14957 (C4 (Web Text))

Max Activating Token Index: 736

Text #14

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.3967. Min Act: -0.0000

Data Index: 1585726 (Python Code)

Max Activating Token Index: 352

Text #15

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.4418. Min Act: -0.0000

Data Index: 1391326 (Python Code)

Max Activating Token Index: 251

Text #16

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.3958. Min Act: -0.0000

Data Index: 485830 (C4 (Web Text))

Max Activating Token Index: 547

Text #17

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.3710. Min Act: 0.0000

Data Index: 1468558 (Python Code)

Max Activating Token Index: 534

Text #18

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.3452. Min Act: -0.0000

Data Index: 1602654 (Python Code)

Max Activating Token Index: 288

Text #19

Max Range: 0.5672. Min Range: -0.5672

Max Act: 0.3683. Min Act: -0.0000

Data Index: 213607 (C4 (Web Text))

Max Activating Token Index: 428