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

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

Neuron 3234 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 3.1183. Min Range: -3.1183

Max Act: 3.1183. Min Act: -0.0001

Data Index: 369720 (C4 (Web Text))

Max Activating Token Index: 144

Text #1

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.9306. Min Act: -0.0001

Data Index: 99615 (C4 (Web Text))

Max Activating Token Index: 329

Text #2

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.7232. Min Act: -0.0000

Data Index: 551353 (C4 (Web Text))

Max Activating Token Index: 337

Text #3

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.1622. Min Act: -0.0001

Data Index: 555533 (C4 (Web Text))

Max Activating Token Index: 212

Text #4

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.1732. Min Act: -0.0001

Data Index: 917704 (C4 (Web Text))

Max Activating Token Index: 922

Text #5

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.9150. Min Act: -0.0000

Data Index: 23016 (C4 (Web Text))

Max Activating Token Index: 660

Text #6

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.0066. Min Act: -0.0000

Data Index: 1112493 (C4 (Web Text))

Max Activating Token Index: 761

Text #7

Max Range: 3.1183. Min Range: -3.1183

Max Act: 2.0011. Min Act: -0.0001

Data Index: 945047 (C4 (Web Text))

Max Activating Token Index: 373

Text #8

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.8406. Min Act: -0.0001

Data Index: 74435 (C4 (Web Text))

Max Activating Token Index: 129

Text #9

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.8251. Min Act: -0.0000

Data Index: 757353 (C4 (Web Text))

Max Activating Token Index: 547

Text #10

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.8075. Min Act: -0.0000

Data Index: 1080443 (C4 (Web Text))

Max Activating Token Index: 893

Text #11

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7496. Min Act: -0.0001

Data Index: 1553465 (Python Code)

Max Activating Token Index: 679

Text #12

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7376. Min Act: -0.0001

Data Index: 1355636 (C4 (Web Text))

Max Activating Token Index: 780

Text #13

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.8064. Min Act: -0.0000

Data Index: 1407976 (Python Code)

Max Activating Token Index: 757

Text #14

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7849. Min Act: -0.0001

Data Index: 862871 (C4 (Web Text))

Max Activating Token Index: 891

Text #15

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7714. Min Act: -0.0001

Data Index: 1565132 (Python Code)

Max Activating Token Index: 487

Text #16

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.6497. Min Act: -0.0000

Data Index: 1237421 (C4 (Web Text))

Max Activating Token Index: 166

Text #17

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7763. Min Act: -0.0001

Data Index: 590305 (C4 (Web Text))

Max Activating Token Index: 811

Text #18

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.7246. Min Act: -0.0001

Data Index: 936246 (C4 (Web Text))

Max Activating Token Index: 908

Text #19

Max Range: 3.1183. Min Range: -3.1183

Max Act: 1.6246. Min Act: -0.0001

Data Index: 172178 (C4 (Web Text))

Max Activating Token Index: 608