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

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

Neuron 650 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8399. Min Act: -0.0001

Data Index: 158458 (C4 (Web Text))

Max Activating Token Index: 736

Text #1

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8153. Min Act: -0.0001

Data Index: 557994 (C4 (Web Text))

Max Activating Token Index: 963

Text #2

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8475. Min Act: -0.0001

Data Index: 1236643 (C4 (Web Text))

Max Activating Token Index: 569

Text #3

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7523. Min Act: -0.0001

Data Index: 363024 (C4 (Web Text))

Max Activating Token Index: 745

Text #4

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7765. Min Act: -0.0001

Data Index: 510419 (C4 (Web Text))

Max Activating Token Index: 764

Text #5

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7950. Min Act: -0.0001

Data Index: 778551 (C4 (Web Text))

Max Activating Token Index: 824

Text #6

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8580. Min Act: -0.0001

Data Index: 951874 (C4 (Web Text))

Max Activating Token Index: 703

Text #7

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8985. Min Act: -0.0001

Data Index: 484474 (C4 (Web Text))

Max Activating Token Index: 1000

Text #8

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8203. Min Act: -0.0001

Data Index: 823026 (C4 (Web Text))

Max Activating Token Index: 396

Text #9

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7417. Min Act: -0.0001

Data Index: 718919 (C4 (Web Text))

Max Activating Token Index: 215

Text #10

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7526. Min Act: -0.0001

Data Index: 659651 (C4 (Web Text))

Max Activating Token Index: 904

Text #11

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.8156. Min Act: -0.0001

Data Index: 552348 (C4 (Web Text))

Max Activating Token Index: 969

Text #12

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7903. Min Act: -0.0001

Data Index: 453810 (C4 (Web Text))

Max Activating Token Index: 380

Text #13

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7451. Min Act: -0.0001

Data Index: 375115 (C4 (Web Text))

Max Activating Token Index: 88

Text #14

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7395. Min Act: -0.0001

Data Index: 938139 (C4 (Web Text))

Max Activating Token Index: 924

Text #15

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7014. Min Act: -0.0001

Data Index: 1224924 (C4 (Web Text))

Max Activating Token Index: 465

Text #16

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7532. Min Act: -0.0001

Data Index: 647589 (C4 (Web Text))

Max Activating Token Index: 996

Text #17

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7408. Min Act: -0.0001

Data Index: 1091913 (C4 (Web Text))

Max Activating Token Index: 987

Text #18

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.6860. Min Act: -0.0001

Data Index: 509017 (C4 (Web Text))

Max Activating Token Index: 304

Text #19

Max Range: 0.8985. Min Range: -0.8985

Max Act: 0.7403. Min Act: -0.0001

Data Index: 234085 (C4 (Web Text))

Max Activating Token Index: 844