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

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

Neuron 563 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: 0.9726. Min Range: -0.9726

Max Act: 0.9726. Min Act: -0.0001

Data Index: 390923 (C4 (Web Text))

Max Activating Token Index: 812

Text #1

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.7915. Min Act: -0.0001

Data Index: 1014247 (C4 (Web Text))

Max Activating Token Index: 300

Text #2

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.8022. Min Act: -0.0001

Data Index: 228114 (C4 (Web Text))

Max Activating Token Index: 875

Text #3

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.8528. Min Act: -0.0001

Data Index: 1016320 (C4 (Web Text))

Max Activating Token Index: 130

Text #4

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.8009. Min Act: -0.0001

Data Index: 995442 (C4 (Web Text))

Max Activating Token Index: 944

Text #5

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.7287. Min Act: -0.0001

Data Index: 67905 (C4 (Web Text))

Max Activating Token Index: 900

Text #6

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.7290. Min Act: -0.0001

Data Index: 7059 (C4 (Web Text))

Max Activating Token Index: 598

Text #7

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.5945. Min Act: -0.0001

Data Index: 282540 (C4 (Web Text))

Max Activating Token Index: 101

Text #8

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6575. Min Act: -0.0001

Data Index: 691678 (C4 (Web Text))

Max Activating Token Index: 661

Text #9

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6349. Min Act: -0.0001

Data Index: 1500864 (Python Code)

Max Activating Token Index: 853

Text #10

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.7606. Min Act: -0.0001

Data Index: 684119 (C4 (Web Text))

Max Activating Token Index: 866

Text #11

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6180. Min Act: -0.0001

Data Index: 286369 (C4 (Web Text))

Max Activating Token Index: 687

Text #12

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.5969. Min Act: -0.0001

Data Index: 475985 (C4 (Web Text))

Max Activating Token Index: 262

Text #13

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6384. Min Act: -0.0001

Data Index: 676328 (C4 (Web Text))

Max Activating Token Index: 209

Text #14

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6564. Min Act: -0.0001

Data Index: 204311 (C4 (Web Text))

Max Activating Token Index: 435

Text #15

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6170. Min Act: -0.0001

Data Index: 963369 (C4 (Web Text))

Max Activating Token Index: 892

Text #16

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6731. Min Act: -0.0001

Data Index: 1061749 (C4 (Web Text))

Max Activating Token Index: 994

Text #17

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6411. Min Act: -0.0001

Data Index: 1341003 (C4 (Web Text))

Max Activating Token Index: 847

Text #18

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.5894. Min Act: -0.0001

Data Index: 1621218 (Python Code)

Max Activating Token Index: 828

Text #19

Max Range: 0.9726. Min Range: -0.9726

Max Act: 0.6313. Min Act: -0.0001

Data Index: 83796 (C4 (Web Text))

Max Activating Token Index: 925