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

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

Neuron 787 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.0009. Min Range: -1.0009

Max Act: 1.0009. Min Act: -0.0000

Data Index: 334183 (C4 (Web Text))

Max Activating Token Index: 72

Text #1

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.9010. Min Act: -0.0000

Data Index: 873176 (C4 (Web Text))

Max Activating Token Index: 71

Text #2

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.8185. Min Act: -0.0000

Data Index: 1143120 (C4 (Web Text))

Max Activating Token Index: 656

Text #3

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.8364. Min Act: -0.0001

Data Index: 412751 (C4 (Web Text))

Max Activating Token Index: 983

Text #4

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.8526. Min Act: -0.0000

Data Index: 484713 (C4 (Web Text))

Max Activating Token Index: 65

Text #5

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.8461. Min Act: -0.0001

Data Index: 789131 (C4 (Web Text))

Max Activating Token Index: 862

Text #6

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.8185. Min Act: -0.0001

Data Index: 27643 (C4 (Web Text))

Max Activating Token Index: 997

Text #7

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7776. Min Act: -0.0000

Data Index: 1081268 (C4 (Web Text))

Max Activating Token Index: 411

Text #8

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7973. Min Act: -0.0001

Data Index: 973102 (C4 (Web Text))

Max Activating Token Index: 963

Text #9

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7803. Min Act: -0.0000

Data Index: 705198 (C4 (Web Text))

Max Activating Token Index: 1008

Text #10

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7017. Min Act: -0.0000

Data Index: 285207 (C4 (Web Text))

Max Activating Token Index: 595

Text #11

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7641. Min Act: -0.0001

Data Index: 1628210 (Python Code)

Max Activating Token Index: 922

Text #12

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7126. Min Act: -0.0001

Data Index: 899385 (C4 (Web Text))

Max Activating Token Index: 761

Text #13

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6509. Min Act: -0.0001

Data Index: 1529248 (Python Code)

Max Activating Token Index: 744

Text #14

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6059. Min Act: -0.0001

Data Index: 1622244 (Python Code)

Max Activating Token Index: 689

Text #15

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.7366. Min Act: -0.0001

Data Index: 1517279 (Python Code)

Max Activating Token Index: 801

Text #16

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6729. Min Act: -0.0001

Data Index: 1127538 (C4 (Web Text))

Max Activating Token Index: 489

Text #17

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6628. Min Act: -0.0001

Data Index: 1525330 (Python Code)

Max Activating Token Index: 752

Text #18

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6881. Min Act: -0.0001

Data Index: 1378668 (Python Code)

Max Activating Token Index: 750

Text #19

Max Range: 1.0009. Min Range: -1.0009

Max Act: 0.6478. Min Act: -0.0001

Data Index: 1652590 (Python Code)

Max Activating Token Index: 679