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

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

Neuron 3578 in Layer 6

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4966. Min Act: -0.0001

Data Index: 1469171 (Python Code)

Max Activating Token Index: 459

Text #1

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4691. Min Act: -0.0001

Data Index: 965873 (C4 (Web Text))

Max Activating Token Index: 250

Text #2

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4309. Min Act: -0.0001

Data Index: 241421 (C4 (Web Text))

Max Activating Token Index: 622

Text #3

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4566. Min Act: -0.0001

Data Index: 1492931 (Python Code)

Max Activating Token Index: 41

Text #4

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4555. Min Act: -0.0001

Data Index: 1461450 (Python Code)

Max Activating Token Index: 663

Text #5

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4442. Min Act: -0.0001

Data Index: 380428 (C4 (Web Text))

Max Activating Token Index: 983

Text #6

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4442. Min Act: -0.0001

Data Index: 1486988 (Python Code)

Max Activating Token Index: 901

Text #7

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4541. Min Act: -0.0001

Data Index: 1070229 (C4 (Web Text))

Max Activating Token Index: 664

Text #8

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4168. Min Act: -0.0001

Data Index: 954600 (C4 (Web Text))

Max Activating Token Index: 421

Text #9

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4250. Min Act: -0.0001

Data Index: 1412054 (Python Code)

Max Activating Token Index: 655

Text #10

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4054. Min Act: -0.0001

Data Index: 1607555 (Python Code)

Max Activating Token Index: 727

Text #11

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.3942. Min Act: -0.0001

Data Index: 1214667 (C4 (Web Text))

Max Activating Token Index: 438

Text #12

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4314. Min Act: -0.0001

Data Index: 451538 (C4 (Web Text))

Max Activating Token Index: 194

Text #13

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4154. Min Act: -0.0001

Data Index: 1142269 (C4 (Web Text))

Max Activating Token Index: 886

Text #14

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4165. Min Act: -0.0001

Data Index: 1078480 (C4 (Web Text))

Max Activating Token Index: 959

Text #15

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4078. Min Act: -0.0001

Data Index: 1514981 (Python Code)

Max Activating Token Index: 1005

Text #16

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4119. Min Act: -0.0001

Data Index: 300440 (C4 (Web Text))

Max Activating Token Index: 736

Text #17

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.4165. Min Act: -0.0001

Data Index: 1633783 (Python Code)

Max Activating Token Index: 52

Text #18

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.3857. Min Act: -0.0001

Data Index: 1492676 (Python Code)

Max Activating Token Index: 279

Text #19

Max Range: 0.4966. Min Range: -0.4966

Max Act: 0.3605. Min Act: -0.0001

Data Index: 1253088 (C4 (Web Text))

Max Activating Token Index: 857