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

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

Neuron 4702 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.9330. Min Act: -0.0001

Data Index: 1287218 (C4 (Web Text))

Max Activating Token Index: 938

Text #1

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8921. Min Act: -0.0001

Data Index: 1007730 (C4 (Web Text))

Max Activating Token Index: 37

Text #2

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8782. Min Act: -0.0001

Data Index: 727367 (C4 (Web Text))

Max Activating Token Index: 510

Text #3

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.9303. Min Act: -0.0001

Data Index: 688141 (C4 (Web Text))

Max Activating Token Index: 572

Text #4

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8213. Min Act: -0.0001

Data Index: 707917 (C4 (Web Text))

Max Activating Token Index: 834

Text #5

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8441. Min Act: -0.0001

Data Index: 1080858 (C4 (Web Text))

Max Activating Token Index: 99

Text #6

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.9081. Min Act: -0.0001

Data Index: 758593 (C4 (Web Text))

Max Activating Token Index: 367

Text #7

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8568. Min Act: -0.0001

Data Index: 1103709 (C4 (Web Text))

Max Activating Token Index: 219

Text #8

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8217. Min Act: -0.0001

Data Index: 708350 (C4 (Web Text))

Max Activating Token Index: 924

Text #9

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7913. Min Act: -0.0001

Data Index: 981007 (C4 (Web Text))

Max Activating Token Index: 326

Text #10

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8474. Min Act: -0.0001

Data Index: 764591 (C4 (Web Text))

Max Activating Token Index: 652

Text #11

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7912. Min Act: -0.0001

Data Index: 757287 (C4 (Web Text))

Max Activating Token Index: 360

Text #12

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7823. Min Act: -0.0001

Data Index: 74399 (C4 (Web Text))

Max Activating Token Index: 332

Text #13

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8640. Min Act: -0.0001

Data Index: 196368 (C4 (Web Text))

Max Activating Token Index: 740

Text #14

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8040. Min Act: -0.0001

Data Index: 316648 (C4 (Web Text))

Max Activating Token Index: 791

Text #15

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8341. Min Act: -0.0001

Data Index: 7164 (C4 (Web Text))

Max Activating Token Index: 7

Text #16

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.8057. Min Act: -0.0001

Data Index: 1149214 (C4 (Web Text))

Max Activating Token Index: 77

Text #17

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7789. Min Act: -0.0001

Data Index: 158628 (C4 (Web Text))

Max Activating Token Index: 803

Text #18

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7459. Min Act: -0.0001

Data Index: 149222 (C4 (Web Text))

Max Activating Token Index: 698

Text #19

Max Range: 0.9330. Min Range: -0.9330

Max Act: 0.7268. Min Act: -0.0001

Data Index: 860541 (C4 (Web Text))

Max Activating Token Index: 191