Model: SoLU Model: 3 Layers, 2048 Neurons per Layer

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

Neuron 225 in Layer 2

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-3l')

Text #0

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0597. Min Act: -0.0003

Data Index: 843157 (C4 (Web Text))

Max Activating Token Index: 443

Text #1

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0555. Min Act: -0.0003

Data Index: 1331306 (C4 (Web Text))

Max Activating Token Index: 437

Text #2

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0526. Min Act: -0.0003

Data Index: 852984 (C4 (Web Text))

Max Activating Token Index: 874

Text #3

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0501. Min Act: -0.0003

Data Index: 643865 (C4 (Web Text))

Max Activating Token Index: 398

Text #4

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0509. Min Act: -0.0003

Data Index: 645754 (C4 (Web Text))

Max Activating Token Index: 715

Text #5

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0445. Min Act: -0.0003

Data Index: 419517 (C4 (Web Text))

Max Activating Token Index: 894

Text #6

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0466. Min Act: -0.0003

Data Index: 963995 (C4 (Web Text))

Max Activating Token Index: 278

Text #7

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0445. Min Act: -0.0003

Data Index: 33038 (C4 (Web Text))

Max Activating Token Index: 667

Text #8

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0435. Min Act: -0.0003

Data Index: 1013797 (C4 (Web Text))

Max Activating Token Index: 120

Text #9

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0430. Min Act: -0.0003

Data Index: 1026857 (C4 (Web Text))

Max Activating Token Index: 843

Text #10

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0409. Min Act: -0.0003

Data Index: 835724 (C4 (Web Text))

Max Activating Token Index: 661

Text #11

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0402. Min Act: -0.0003

Data Index: 1272783 (C4 (Web Text))

Max Activating Token Index: 288

Text #12

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0410. Min Act: -0.0003

Data Index: 913657 (C4 (Web Text))

Max Activating Token Index: 870

Text #13

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0384. Min Act: -0.0003

Data Index: 835085 (C4 (Web Text))

Max Activating Token Index: 466

Text #14

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0372. Min Act: -0.0003

Data Index: 1571605 (Python Code)

Max Activating Token Index: 588

Text #15

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0374. Min Act: -0.0003

Data Index: 747823 (C4 (Web Text))

Max Activating Token Index: 666

Text #16

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0373. Min Act: -0.0003

Data Index: 746980 (C4 (Web Text))

Max Activating Token Index: 534

Text #17

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0384. Min Act: -0.0003

Data Index: 677547 (C4 (Web Text))

Max Activating Token Index: 189

Text #18

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0391. Min Act: -0.0003

Data Index: 59550 (C4 (Web Text))

Max Activating Token Index: 989

Text #19

Max Range: 0.0597. Min Range: -0.0597

Max Act: 0.0365. Min Act: -0.0003

Data Index: 550786 (C4 (Web Text))

Max Activating Token Index: 988