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

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

Neuron 1620 in Layer 0

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-4l')

Text #0

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0101. Min Act: -0.0005

Data Index: 1403955 (Python Code)

Max Activating Token Index: 335

Text #1

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0094. Min Act: -0.0005

Data Index: 804917 (C4 (Web Text))

Max Activating Token Index: 811

Text #2

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0095. Min Act: -0.0005

Data Index: 810478 (C4 (Web Text))

Max Activating Token Index: 289

Text #3

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0094. Min Act: -0.0004

Data Index: 1126200 (C4 (Web Text))

Max Activating Token Index: 204

Text #4

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0094. Min Act: -0.0005

Data Index: 704655 (C4 (Web Text))

Max Activating Token Index: 788

Text #5

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0091. Min Act: -0.0006

Data Index: 768303 (C4 (Web Text))

Max Activating Token Index: 15

Text #6

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0089. Min Act: -0.0005

Data Index: 947886 (C4 (Web Text))

Max Activating Token Index: 162

Text #7

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0089. Min Act: -0.0005

Data Index: 341147 (C4 (Web Text))

Max Activating Token Index: 285

Text #8

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0090. Min Act: -0.0005

Data Index: 52767 (C4 (Web Text))

Max Activating Token Index: 264

Text #9

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0089. Min Act: -0.0005

Data Index: 1268270 (C4 (Web Text))

Max Activating Token Index: 247

Text #10

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0087. Min Act: -0.0005

Data Index: 1535050 (Python Code)

Max Activating Token Index: 500

Text #11

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0082. Min Act: -0.0006

Data Index: 330365 (C4 (Web Text))

Max Activating Token Index: 723

Text #12

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0084. Min Act: -0.0005

Data Index: 835705 (C4 (Web Text))

Max Activating Token Index: 923

Text #13

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0086. Min Act: -0.0005

Data Index: 1152150 (C4 (Web Text))

Max Activating Token Index: 797

Text #14

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0085. Min Act: -0.0005

Data Index: 1448476 (Python Code)

Max Activating Token Index: 628

Text #15

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0085. Min Act: -0.0005

Data Index: 698454 (C4 (Web Text))

Max Activating Token Index: 34

Text #16

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0085. Min Act: -0.0005

Data Index: 550046 (C4 (Web Text))

Max Activating Token Index: 293

Text #17

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0087. Min Act: -0.0005

Data Index: 1633715 (Python Code)

Max Activating Token Index: 692

Text #18

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0084. Min Act: -0.0005

Data Index: 1165571 (C4 (Web Text))

Max Activating Token Index: 349

Text #19

Max Range: 0.0101. Min Range: -0.0101

Max Act: 0.0083. Min Act: -0.0005

Data Index: 1532536 (Python Code)

Max Activating Token Index: 420