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

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

Neuron 731 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: 0.2407. Min Range: -0.2407

Max Act: 0.2407. Min Act: -0.0001

Data Index: 1515528 (Python Code)

Max Activating Token Index: 611

Text #1

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2332. Min Act: -0.0001

Data Index: 678068 (C4 (Web Text))

Max Activating Token Index: 583

Text #2

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2395. Min Act: -0.0001

Data Index: 1236733 (C4 (Web Text))

Max Activating Token Index: 556

Text #3

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2304. Min Act: -0.0001

Data Index: 425096 (C4 (Web Text))

Max Activating Token Index: 319

Text #4

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2115. Min Act: -0.0001

Data Index: 407544 (C4 (Web Text))

Max Activating Token Index: 1001

Text #5

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2343. Min Act: -0.0001

Data Index: 120186 (C4 (Web Text))

Max Activating Token Index: 769

Text #6

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1978. Min Act: -0.0001

Data Index: 1166019 (C4 (Web Text))

Max Activating Token Index: 929

Text #7

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.2016. Min Act: -0.0001

Data Index: 1413441 (Python Code)

Max Activating Token Index: 848

Text #8

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1751. Min Act: -0.0001

Data Index: 1654987 (Python Code)

Max Activating Token Index: 245

Text #9

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1652. Min Act: -0.0001

Data Index: 1502017 (Python Code)

Max Activating Token Index: 807

Text #10

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1663. Min Act: -0.0000

Data Index: 26071 (C4 (Web Text))

Max Activating Token Index: 404

Text #11

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1699. Min Act: -0.0000

Data Index: 104908 (C4 (Web Text))

Max Activating Token Index: 499

Text #12

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1579. Min Act: -0.0001

Data Index: 1411961 (Python Code)

Max Activating Token Index: 235

Text #13

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1580. Min Act: -0.0001

Data Index: 871916 (C4 (Web Text))

Max Activating Token Index: 416

Text #14

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1456. Min Act: -0.0001

Data Index: 1366344 (Python Code)

Max Activating Token Index: 707

Text #15

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1479. Min Act: -0.0001

Data Index: 260246 (C4 (Web Text))

Max Activating Token Index: 795

Text #16

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1527. Min Act: -0.0001

Data Index: 1468407 (Python Code)

Max Activating Token Index: 107

Text #17

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1469. Min Act: -0.0001

Data Index: 1607262 (Python Code)

Max Activating Token Index: 731

Text #18

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1598. Min Act: -0.0001

Data Index: 1489867 (Python Code)

Max Activating Token Index: 501

Text #19

Max Range: 0.2407. Min Range: -0.2407

Max Act: 0.1425. Min Act: -0.0001

Data Index: 1506339 (Python Code)

Max Activating Token Index: 980