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

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

Neuron 3159 in Layer 5

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1406. Min Act: -0.0001

Data Index: 850091 (C4 (Web Text))

Max Activating Token Index: 554

Text #1

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1520. Min Act: -0.0000

Data Index: 1647516 (Python Code)

Max Activating Token Index: 699

Text #2

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1434. Min Act: -0.0000

Data Index: 212510 (C4 (Web Text))

Max Activating Token Index: 180

Text #3

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1485. Min Act: -0.0000

Data Index: 1032908 (C4 (Web Text))

Max Activating Token Index: 674

Text #4

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1462. Min Act: -0.0001

Data Index: 5512 (C4 (Web Text))

Max Activating Token Index: 435

Text #5

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1408. Min Act: -0.0000

Data Index: 1251505 (C4 (Web Text))

Max Activating Token Index: 852

Text #6

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1464. Min Act: -0.0000

Data Index: 1142947 (C4 (Web Text))

Max Activating Token Index: 615

Text #7

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1399. Min Act: -0.0000

Data Index: 420203 (C4 (Web Text))

Max Activating Token Index: 404

Text #8

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1403. Min Act: -0.0000

Data Index: 733973 (C4 (Web Text))

Max Activating Token Index: 404

Text #9

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1340. Min Act: -0.0001

Data Index: 34359 (C4 (Web Text))

Max Activating Token Index: 118

Text #10

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1325. Min Act: -0.0001

Data Index: 1447135 (Python Code)

Max Activating Token Index: 14

Text #11

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1306. Min Act: -0.0000

Data Index: 474221 (C4 (Web Text))

Max Activating Token Index: 894

Text #12

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1202. Min Act: -0.0000

Data Index: 243088 (C4 (Web Text))

Max Activating Token Index: 128

Text #13

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1383. Min Act: -0.0000

Data Index: 868666 (C4 (Web Text))

Max Activating Token Index: 695

Text #14

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1159. Min Act: -0.0001

Data Index: 1494183 (Python Code)

Max Activating Token Index: 386

Text #15

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1224. Min Act: -0.0001

Data Index: 75936 (C4 (Web Text))

Max Activating Token Index: 552

Text #16

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1260. Min Act: -0.0001

Data Index: 1535200 (Python Code)

Max Activating Token Index: 719

Text #17

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1189. Min Act: -0.0001

Data Index: 1633881 (Python Code)

Max Activating Token Index: 505

Text #18

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1212. Min Act: -0.0000

Data Index: 1291444 (C4 (Web Text))

Max Activating Token Index: 111

Text #19

Max Range: 0.1520. Min Range: -0.1520

Max Act: 0.1230. Min Act: -0.0000

Data Index: 167703 (C4 (Web Text))

Max Activating Token Index: 777