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

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

Neuron 3218 in Layer 8

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.1198. Min Act: -0.0000

Data Index: 791284 (C4 (Web Text))

Max Activating Token Index: 306

Text #1

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.0929. Min Act: -0.0000

Data Index: 1547180 (Python Code)

Max Activating Token Index: 589

Text #2

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.0064. Min Act: -0.0000

Data Index: 586095 (C4 (Web Text))

Max Activating Token Index: 54

Text #3

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.0182. Min Act: -0.0000

Data Index: 1591459 (Python Code)

Max Activating Token Index: 387

Text #4

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.0082. Min Act: -0.0000

Data Index: 1609322 (Python Code)

Max Activating Token Index: 616

Text #5

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9463. Min Act: -0.0000

Data Index: 1297530 (C4 (Web Text))

Max Activating Token Index: 361

Text #6

Max Range: 1.1198. Min Range: -1.1198

Max Act: 1.0278. Min Act: -0.0000

Data Index: 87599 (C4 (Web Text))

Max Activating Token Index: 208

Text #7

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9151. Min Act: -0.0000

Data Index: 221100 (C4 (Web Text))

Max Activating Token Index: 894

Text #8

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9270. Min Act: -0.0000

Data Index: 889745 (C4 (Web Text))

Max Activating Token Index: 177

Text #9

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9118. Min Act: -0.0000

Data Index: 607336 (C4 (Web Text))

Max Activating Token Index: 898

Text #10

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9533. Min Act: -0.0000

Data Index: 745424 (C4 (Web Text))

Max Activating Token Index: 698

Text #11

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9789. Min Act: -0.0001

Data Index: 664142 (C4 (Web Text))

Max Activating Token Index: 383

Text #12

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9501. Min Act: -0.0000

Data Index: 988504 (C4 (Web Text))

Max Activating Token Index: 868

Text #13

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9291. Min Act: -0.0000

Data Index: 39173 (C4 (Web Text))

Max Activating Token Index: 86

Text #14

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9353. Min Act: -0.0000

Data Index: 1361772 (Python Code)

Max Activating Token Index: 342

Text #15

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.9037. Min Act: -0.0000

Data Index: 388778 (C4 (Web Text))

Max Activating Token Index: 536

Text #16

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.8661. Min Act: -0.0000

Data Index: 630312 (C4 (Web Text))

Max Activating Token Index: 796

Text #17

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.8342. Min Act: -0.0000

Data Index: 250016 (C4 (Web Text))

Max Activating Token Index: 622

Text #18

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.8852. Min Act: -0.0000

Data Index: 269099 (C4 (Web Text))

Max Activating Token Index: 263

Text #19

Max Range: 1.1198. Min Range: -1.1198

Max Act: 0.8545. Min Act: -0.0001

Data Index: 62330 (C4 (Web Text))

Max Activating Token Index: 974