Model: SoLU Model: 6 Layers, 3072 Neurons per Layer

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

Neuron 302 in Layer 4

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-6l')

Text #0

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.8634. Min Act: -0.0001

Data Index: 1197877 (C4 (Web Text))

Max Activating Token Index: 299

Text #1

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7957. Min Act: -0.0001

Data Index: 1079034 (C4 (Web Text))

Max Activating Token Index: 458

Text #2

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.8230. Min Act: -0.0001

Data Index: 1640996 (Python Code)

Max Activating Token Index: 914

Text #3

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7874. Min Act: -0.0001

Data Index: 1455795 (Python Code)

Max Activating Token Index: 656

Text #4

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7696. Min Act: -0.0001

Data Index: 932519 (C4 (Web Text))

Max Activating Token Index: 162

Text #5

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.8009. Min Act: -0.0001

Data Index: 1561910 (Python Code)

Max Activating Token Index: 940

Text #6

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7238. Min Act: -0.0001

Data Index: 619399 (C4 (Web Text))

Max Activating Token Index: 834

Text #7

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7450. Min Act: -0.0001

Data Index: 995056 (C4 (Web Text))

Max Activating Token Index: 933

Text #8

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7048. Min Act: -0.0001

Data Index: 1029843 (C4 (Web Text))

Max Activating Token Index: 621

Text #9

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.6624. Min Act: -0.0001

Data Index: 1398233 (Python Code)

Max Activating Token Index: 857

Text #10

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7053. Min Act: -0.0001

Data Index: 470795 (C4 (Web Text))

Max Activating Token Index: 406

Text #11

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7329. Min Act: -0.0001

Data Index: 1432881 (Python Code)

Max Activating Token Index: 375

Text #12

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7908. Min Act: -0.0001

Data Index: 1602424 (Python Code)

Max Activating Token Index: 275

Text #13

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.6968. Min Act: -0.0001

Data Index: 121261 (C4 (Web Text))

Max Activating Token Index: 151

Text #14

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7399. Min Act: -0.0001

Data Index: 998948 (C4 (Web Text))

Max Activating Token Index: 283

Text #15

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.6588. Min Act: -0.0001

Data Index: 838190 (C4 (Web Text))

Max Activating Token Index: 389

Text #16

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7173. Min Act: -0.0001

Data Index: 1596077 (Python Code)

Max Activating Token Index: 863

Text #17

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.6805. Min Act: -0.0001

Data Index: 443417 (C4 (Web Text))

Max Activating Token Index: 204

Text #18

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.6886. Min Act: -0.0001

Data Index: 939982 (C4 (Web Text))

Max Activating Token Index: 132

Text #19

Max Range: 0.8634. Min Range: -0.8634

Max Act: 0.7015. Min Act: -0.0001

Data Index: 785054 (C4 (Web Text))

Max Activating Token Index: 885