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

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

Neuron 1113 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.8088. Min Act: -0.0001

Data Index: 606650 (C4 (Web Text))

Max Activating Token Index: 677

Text #1

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.7423. Min Act: -0.0001

Data Index: 109374 (C4 (Web Text))

Max Activating Token Index: 257

Text #2

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6846. Min Act: -0.0001

Data Index: 1268815 (C4 (Web Text))

Max Activating Token Index: 91

Text #3

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6990. Min Act: -0.0001

Data Index: 719536 (C4 (Web Text))

Max Activating Token Index: 332

Text #4

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6752. Min Act: -0.0001

Data Index: 1145894 (C4 (Web Text))

Max Activating Token Index: 293

Text #5

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6671. Min Act: -0.0001

Data Index: 426345 (C4 (Web Text))

Max Activating Token Index: 450

Text #6

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6164. Min Act: -0.0001

Data Index: 792888 (C4 (Web Text))

Max Activating Token Index: 80

Text #7

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6585. Min Act: -0.0001

Data Index: 1620964 (Python Code)

Max Activating Token Index: 465

Text #8

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6548. Min Act: -0.0001

Data Index: 687850 (C4 (Web Text))

Max Activating Token Index: 601

Text #9

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.5827. Min Act: -0.0001

Data Index: 1442199 (Python Code)

Max Activating Token Index: 1021

Text #10

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6437. Min Act: -0.0001

Data Index: 1515144 (Python Code)

Max Activating Token Index: 686

Text #11

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6362. Min Act: -0.0001

Data Index: 1473430 (Python Code)

Max Activating Token Index: 424

Text #12

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6363. Min Act: -0.0001

Data Index: 452328 (C4 (Web Text))

Max Activating Token Index: 410

Text #13

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6579. Min Act: -0.0001

Data Index: 645075 (C4 (Web Text))

Max Activating Token Index: 909

Text #14

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6779. Min Act: -0.0001

Data Index: 1360518 (Python Code)

Max Activating Token Index: 605

Text #15

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.7044. Min Act: -0.0001

Data Index: 1042228 (C4 (Web Text))

Max Activating Token Index: 537

Text #16

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6575. Min Act: -0.0001

Data Index: 1599777 (Python Code)

Max Activating Token Index: 323

Text #17

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6428. Min Act: -0.0001

Data Index: 1085430 (C4 (Web Text))

Max Activating Token Index: 786

Text #18

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6033. Min Act: -0.0001

Data Index: 1415754 (Python Code)

Max Activating Token Index: 454

Text #19

Max Range: 0.8088. Min Range: -0.8088

Max Act: 0.6420. Min Act: -0.0001

Data Index: 511127 (C4 (Web Text))

Max Activating Token Index: 322