Model: SoLU Model: 10 Layers, 5120 Neurons per Layer

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

Neuron 1088 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-10l')

Text #0

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.4764. Min Act: -0.0001

Data Index: 900205 (C4 (Web Text))

Max Activating Token Index: 365

Text #1

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.4935. Min Act: -0.0001

Data Index: 697329 (C4 (Web Text))

Max Activating Token Index: 360

Text #2

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.4267. Min Act: -0.0001

Data Index: 979299 (C4 (Web Text))

Max Activating Token Index: 464

Text #3

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.1950. Min Act: -0.0001

Data Index: 1499743 (Python Code)

Max Activating Token Index: 1009

Text #4

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2919. Min Act: -0.0001

Data Index: 1327992 (C4 (Web Text))

Max Activating Token Index: 942

Text #5

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2665. Min Act: -0.0001

Data Index: 847927 (C4 (Web Text))

Max Activating Token Index: 642

Text #6

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2033. Min Act: -0.0001

Data Index: 625406 (C4 (Web Text))

Max Activating Token Index: 639

Text #7

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2990. Min Act: -0.0001

Data Index: 117492 (C4 (Web Text))

Max Activating Token Index: 736

Text #8

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2865. Min Act: -0.0001

Data Index: 774852 (C4 (Web Text))

Max Activating Token Index: 57

Text #9

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.3570. Min Act: -0.0001

Data Index: 626933 (C4 (Web Text))

Max Activating Token Index: 223

Text #10

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2663. Min Act: -0.0001

Data Index: 21483 (C4 (Web Text))

Max Activating Token Index: 108

Text #11

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.3082. Min Act: -0.0001

Data Index: 899796 (C4 (Web Text))

Max Activating Token Index: 571

Text #12

Max Range: 1.4935. Min Range: -1.4935

Max Act: 0.9806. Min Act: -0.0001

Data Index: 1293228 (C4 (Web Text))

Max Activating Token Index: 250

Text #13

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2549. Min Act: -0.0001

Data Index: 450786 (C4 (Web Text))

Max Activating Token Index: 473

Text #14

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.3861. Min Act: -0.0001

Data Index: 1386533 (Python Code)

Max Activating Token Index: 331

Text #15

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.0835. Min Act: -0.0001

Data Index: 660300 (C4 (Web Text))

Max Activating Token Index: 300

Text #16

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.0911. Min Act: -0.0001

Data Index: 266035 (C4 (Web Text))

Max Activating Token Index: 171

Text #17

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.2472. Min Act: -0.0001

Data Index: 1622080 (Python Code)

Max Activating Token Index: 926

Text #18

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.1501. Min Act: -0.0001

Data Index: 962532 (C4 (Web Text))

Max Activating Token Index: 235

Text #19

Max Range: 1.4935. Min Range: -1.4935

Max Act: 1.3068. Min Act: -0.0001

Data Index: 1352179 (C4 (Web Text))

Max Activating Token Index: 519