Model: SoLU Model: 12 Layers, 6144 Neurons per Layer

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

Neuron 2835 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

Transformer Lens Loading: HookedTransformer.from_pretrained('solu-12l')

Text #0

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.5383. Min Act: -0.0000

Data Index: 1651149 (Python Code)

Max Activating Token Index: 801

Text #1

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.5916. Min Act: -0.0000

Data Index: 206544 (C4 (Web Text))

Max Activating Token Index: 907

Text #2

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.3462. Min Act: -0.0000

Data Index: 711461 (C4 (Web Text))

Max Activating Token Index: 630

Text #3

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.2987. Min Act: -0.0000

Data Index: 334092 (C4 (Web Text))

Max Activating Token Index: 934

Text #4

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.3805. Min Act: -0.0000

Data Index: 877009 (C4 (Web Text))

Max Activating Token Index: 947

Text #5

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1214. Min Act: -0.0000

Data Index: 330890 (C4 (Web Text))

Max Activating Token Index: 946

Text #6

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1932. Min Act: -0.0000

Data Index: 167172 (C4 (Web Text))

Max Activating Token Index: 558

Text #7

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1881. Min Act: -0.0000

Data Index: 881409 (C4 (Web Text))

Max Activating Token Index: 200

Text #8

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1302. Min Act: -0.0000

Data Index: 536356 (C4 (Web Text))

Max Activating Token Index: 633

Text #9

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1023. Min Act: -0.0000

Data Index: 307486 (C4 (Web Text))

Max Activating Token Index: 432

Text #10

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.9688. Min Act: -0.0000

Data Index: 931527 (C4 (Web Text))

Max Activating Token Index: 416

Text #11

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1846. Min Act: -0.0000

Data Index: 1100508 (C4 (Web Text))

Max Activating Token Index: 365

Text #12

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.0450. Min Act: -0.0000

Data Index: 1608164 (Python Code)

Max Activating Token Index: 985

Text #13

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.0277. Min Act: -0.0000

Data Index: 278053 (C4 (Web Text))

Max Activating Token Index: 746

Text #14

Max Range: 2.5916. Min Range: -2.5916

Max Act: 2.1284. Min Act: -0.0000

Data Index: 481707 (C4 (Web Text))

Max Activating Token Index: 818

Text #15

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.9921. Min Act: -0.0000

Data Index: 1287457 (C4 (Web Text))

Max Activating Token Index: 388

Text #16

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.9871. Min Act: -0.0000

Data Index: 280867 (C4 (Web Text))

Max Activating Token Index: 368

Text #17

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.8898. Min Act: -0.0000

Data Index: 1266756 (C4 (Web Text))

Max Activating Token Index: 331

Text #18

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.8846. Min Act: -0.0000

Data Index: 728080 (C4 (Web Text))

Max Activating Token Index: 652

Text #19

Max Range: 2.5916. Min Range: -2.5916

Max Act: 1.9942. Min Act: -0.0000

Data Index: 1015314 (C4 (Web Text))

Max Activating Token Index: 13