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

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

Neuron 3883 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.9506. Min Act: -0.0001

Data Index: 1068355 (C4 (Web Text))

Max Activating Token Index: 514

Text #1

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.8401. Min Act: -0.0001

Data Index: 86278 (C4 (Web Text))

Max Activating Token Index: 394

Text #2

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.8218. Min Act: -0.0001

Data Index: 697775 (C4 (Web Text))

Max Activating Token Index: 940

Text #3

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.7571. Min Act: -0.0001

Data Index: 1353800 (C4 (Web Text))

Max Activating Token Index: 321

Text #4

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.8268. Min Act: -0.0001

Data Index: 1010588 (C4 (Web Text))

Max Activating Token Index: 395

Text #5

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.5705. Min Act: -0.0001

Data Index: 849995 (C4 (Web Text))

Max Activating Token Index: 877

Text #6

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.6407. Min Act: -0.0001

Data Index: 1254953 (C4 (Web Text))

Max Activating Token Index: 133

Text #7

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.4820. Min Act: -0.0001

Data Index: 1173319 (C4 (Web Text))

Max Activating Token Index: 921

Text #8

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.5009. Min Act: -0.0001

Data Index: 1633892 (Python Code)

Max Activating Token Index: 31

Text #9

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.4004. Min Act: -0.0001

Data Index: 1380647 (Python Code)

Max Activating Token Index: 24

Text #10

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.3559. Min Act: -0.0001

Data Index: 1589116 (Python Code)

Max Activating Token Index: 37

Text #11

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.3798. Min Act: -0.0001

Data Index: 1594958 (Python Code)

Max Activating Token Index: 62

Text #12

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2649. Min Act: -0.0001

Data Index: 492839 (C4 (Web Text))

Max Activating Token Index: 597

Text #13

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2837. Min Act: -0.0001

Data Index: 1367331 (Python Code)

Max Activating Token Index: 21

Text #14

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2353. Min Act: -0.0001

Data Index: 1362053 (Python Code)

Max Activating Token Index: 811

Text #15

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.1884. Min Act: -0.0001

Data Index: 1546077 (Python Code)

Max Activating Token Index: 55

Text #16

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.3326. Min Act: -0.0001

Data Index: 1573953 (Python Code)

Max Activating Token Index: 551

Text #17

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2601. Min Act: -0.0001

Data Index: 1443930 (Python Code)

Max Activating Token Index: 64

Text #18

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2600. Min Act: -0.0001

Data Index: 1391926 (Python Code)

Max Activating Token Index: 32

Text #19

Max Range: 1.9506. Min Range: -1.9506

Max Act: 1.2064. Min Act: -0.0001

Data Index: 784523 (C4 (Web Text))

Max Activating Token Index: 355