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

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

Neuron 3977 in Layer 3

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4376. Min Act: -0.0001

Data Index: 258454 (C4 (Web Text))

Max Activating Token Index: 887

Text #1

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4246. Min Act: -0.0001

Data Index: 246943 (C4 (Web Text))

Max Activating Token Index: 976

Text #2

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4479. Min Act: -0.0001

Data Index: 914624 (C4 (Web Text))

Max Activating Token Index: 11

Text #3

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4428. Min Act: -0.0001

Data Index: 468320 (C4 (Web Text))

Max Activating Token Index: 343

Text #4

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4169. Min Act: -0.0001

Data Index: 841295 (C4 (Web Text))

Max Activating Token Index: 368

Text #5

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3979. Min Act: -0.0001

Data Index: 270327 (C4 (Web Text))

Max Activating Token Index: 155

Text #6

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.4124. Min Act: -0.0001

Data Index: 1491436 (Python Code)

Max Activating Token Index: 627

Text #7

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3599. Min Act: -0.0001

Data Index: 892088 (C4 (Web Text))

Max Activating Token Index: 267

Text #8

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3442. Min Act: -0.0001

Data Index: 181491 (C4 (Web Text))

Max Activating Token Index: 310

Text #9

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3303. Min Act: -0.0001

Data Index: 1651618 (Python Code)

Max Activating Token Index: 641

Text #10

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3681. Min Act: -0.0001

Data Index: 1615312 (Python Code)

Max Activating Token Index: 954

Text #11

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3277. Min Act: -0.0001

Data Index: 1534630 (Python Code)

Max Activating Token Index: 68

Text #12

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3501. Min Act: -0.0001

Data Index: 407738 (C4 (Web Text))

Max Activating Token Index: 704

Text #13

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3455. Min Act: -0.0001

Data Index: 504185 (C4 (Web Text))

Max Activating Token Index: 18

Text #14

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3333. Min Act: -0.0001

Data Index: 1618656 (Python Code)

Max Activating Token Index: 987

Text #15

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3265. Min Act: -0.0001

Data Index: 275425 (C4 (Web Text))

Max Activating Token Index: 367

Text #16

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3543. Min Act: -0.0001

Data Index: 6880 (C4 (Web Text))

Max Activating Token Index: 221

Text #17

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3422. Min Act: -0.0001

Data Index: 503972 (C4 (Web Text))

Max Activating Token Index: 911

Text #18

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3261. Min Act: -0.0001

Data Index: 1055586 (C4 (Web Text))

Max Activating Token Index: 299

Text #19

Max Range: 0.4479. Min Range: -0.4479

Max Act: 0.3026. Min Act: -0.0001

Data Index: 640728 (C4 (Web Text))

Max Activating Token Index: 839