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

Dataset: The Pile

Neuron 4791 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 9.1611. Min Range: -9.1611

Max Act: 9.1611. Min Act: -0.0000

Data Index: 1363023 (The Pile)

Max Activating Token Index: 698

Text #1

Max Range: 9.1611. Min Range: -9.1611

Max Act: 8.2693. Min Act: -0.0000

Data Index: 669055 (The Pile)

Max Activating Token Index: 1005

Text #2

Max Range: 9.1611. Min Range: -9.1611

Max Act: 8.4967. Min Act: -0.0000

Data Index: 1191361 (The Pile)

Max Activating Token Index: 853

Text #3

Max Range: 9.1611. Min Range: -9.1611

Max Act: 8.4969. Min Act: -0.0000

Data Index: 174244 (The Pile)

Max Activating Token Index: 541

Text #4

Max Range: 9.1611. Min Range: -9.1611

Max Act: 8.1358. Min Act: -0.0000

Data Index: 1369667 (The Pile)

Max Activating Token Index: 686

Text #5

Max Range: 9.1611. Min Range: -9.1611

Max Act: 7.5885. Min Act: -0.0000

Data Index: 1518430 (The Pile)

Max Activating Token Index: 690

Text #6

Max Range: 9.1611. Min Range: -9.1611

Max Act: 7.1208. Min Act: -0.0000

Data Index: 818797 (The Pile)

Max Activating Token Index: 522

Text #7

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.8240. Min Act: -0.0000

Data Index: 267100 (The Pile)

Max Activating Token Index: 1021

Text #8

Max Range: 9.1611. Min Range: -9.1611

Max Act: 7.0624. Min Act: -0.0000

Data Index: 1059960 (The Pile)

Max Activating Token Index: 756

Text #9

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.7806. Min Act: -0.0000

Data Index: 1985041 (The Pile)

Max Activating Token Index: 955

Text #10

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.8200. Min Act: -0.0000

Data Index: 1709095 (The Pile)

Max Activating Token Index: 778

Text #11

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.5348. Min Act: -0.0000

Data Index: 1364578 (The Pile)

Max Activating Token Index: 436

Text #12

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.3228. Min Act: -0.0000

Data Index: 918740 (The Pile)

Max Activating Token Index: 727

Text #13

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.4247. Min Act: -0.0000

Data Index: 152547 (The Pile)

Max Activating Token Index: 1019

Text #14

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.2146. Min Act: -0.0000

Data Index: 595338 (The Pile)

Max Activating Token Index: 765

Text #15

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.0529. Min Act: -0.0000

Data Index: 1882207 (The Pile)

Max Activating Token Index: 400

Text #16

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.1891. Min Act: -0.0000

Data Index: 465077 (The Pile)

Max Activating Token Index: 512

Text #17

Max Range: 9.1611. Min Range: -9.1611

Max Act: 5.7889. Min Act: -0.0000

Data Index: 811666 (The Pile)

Max Activating Token Index: 522

Text #18

Max Range: 9.1611. Min Range: -9.1611

Max Act: 5.7349. Min Act: -0.0000

Data Index: 139406 (The Pile)

Max Activating Token Index: 315

Text #19

Max Range: 9.1611. Min Range: -9.1611

Max Act: 6.0930. Min Act: -0.0000

Data Index: 1672679 (The Pile)

Max Activating Token Index: 304