Model: SoLU Model: 8 Layers, 4096 Neurons per Layer

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

Neuron 721 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-8l')

Text #0

Max Range: 1.0247. Min Range: -1.0247

Max Act: 1.0247. Min Act: -0.0001

Data Index: 536402 (C4 (Web Text))

Max Activating Token Index: 248

Text #1

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.8396. Min Act: -0.0001

Data Index: 1646104 (Python Code)

Max Activating Token Index: 742

Text #2

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.7377. Min Act: -0.0001

Data Index: 646390 (C4 (Web Text))

Max Activating Token Index: 995

Text #3

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.7591. Min Act: -0.0001

Data Index: 185405 (C4 (Web Text))

Max Activating Token Index: 12

Text #4

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.7084. Min Act: -0.0001

Data Index: 728897 (C4 (Web Text))

Max Activating Token Index: 126

Text #5

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.7289. Min Act: -0.0001

Data Index: 1504069 (Python Code)

Max Activating Token Index: 317

Text #6

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6921. Min Act: -0.0001

Data Index: 972008 (C4 (Web Text))

Max Activating Token Index: 198

Text #7

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6940. Min Act: -0.0001

Data Index: 1590129 (Python Code)

Max Activating Token Index: 440

Text #8

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6919. Min Act: -0.0001

Data Index: 1060117 (C4 (Web Text))

Max Activating Token Index: 24

Text #9

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6440. Min Act: -0.0001

Data Index: 1490879 (Python Code)

Max Activating Token Index: 511

Text #10

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6577. Min Act: -0.0001

Data Index: 161952 (C4 (Web Text))

Max Activating Token Index: 517

Text #11

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6536. Min Act: -0.0001

Data Index: 961708 (C4 (Web Text))

Max Activating Token Index: 192

Text #12

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6634. Min Act: -0.0001

Data Index: 756274 (C4 (Web Text))

Max Activating Token Index: 955

Text #13

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6608. Min Act: -0.0001

Data Index: 1643596 (Python Code)

Max Activating Token Index: 12

Text #14

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6651. Min Act: -0.0001

Data Index: 1387399 (Python Code)

Max Activating Token Index: 176

Text #15

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.5930. Min Act: -0.0001

Data Index: 1602722 (Python Code)

Max Activating Token Index: 929

Text #16

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6521. Min Act: -0.0001

Data Index: 965292 (C4 (Web Text))

Max Activating Token Index: 111

Text #17

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6430. Min Act: -0.0001

Data Index: 646921 (C4 (Web Text))

Max Activating Token Index: 158

Text #18

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6436. Min Act: -0.0001

Data Index: 482228 (C4 (Web Text))

Max Activating Token Index: 260

Text #19

Max Range: 1.0247. Min Range: -1.0247

Max Act: 0.6466. Min Act: -0.0001

Data Index: 990435 (C4 (Web Text))

Max Activating Token Index: 1004