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

Dataset: The Pile

Neuron 3021 in Layer 7

Load this data into an Interactive Neuroscope

See Documentation here

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

Text #0

Max Range: 5.1993. Min Range: -5.1993

Max Act: 5.1993. Min Act: -0.0000

Data Index: 365895 (The Pile)

Max Activating Token Index: 859

Text #1

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.9495. Min Act: -0.0000

Data Index: 799883 (The Pile)

Max Activating Token Index: 810

Text #2

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.7774. Min Act: -0.0000

Data Index: 152520 (The Pile)

Max Activating Token Index: 203

Text #3

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.5450. Min Act: -0.0000

Data Index: 661618 (The Pile)

Max Activating Token Index: 493

Text #4

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.5289. Min Act: -0.0000

Data Index: 101062 (The Pile)

Max Activating Token Index: 984

Text #5

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.5579. Min Act: -0.0000

Data Index: 131263 (The Pile)

Max Activating Token Index: 776

Text #6

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.1504. Min Act: -0.0000

Data Index: 1278866 (The Pile)

Max Activating Token Index: 155

Text #7

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.3988. Min Act: -0.0000

Data Index: 1588080 (The Pile)

Max Activating Token Index: 100

Text #8

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.4195. Min Act: -0.0000

Data Index: 1642169 (The Pile)

Max Activating Token Index: 88

Text #9

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.3412. Min Act: -0.0000

Data Index: 1340083 (The Pile)

Max Activating Token Index: 938

Text #10

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.2485. Min Act: -0.0000

Data Index: 1433793 (The Pile)

Max Activating Token Index: 346

Text #11

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.8583. Min Act: -0.0000

Data Index: 661047 (The Pile)

Max Activating Token Index: 484

Text #12

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.8035. Min Act: -0.0000

Data Index: 1209570 (The Pile)

Max Activating Token Index: 226

Text #13

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.3942. Min Act: -0.0000

Data Index: 1526496 (The Pile)

Max Activating Token Index: 793

Text #14

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.9924. Min Act: -0.0000

Data Index: 624423 (The Pile)

Max Activating Token Index: 475

Text #15

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.1404. Min Act: -0.0000

Data Index: 1690134 (The Pile)

Max Activating Token Index: 448

Text #16

Max Range: 5.1993. Min Range: -5.1993

Max Act: 4.1774. Min Act: -0.0000

Data Index: 720875 (The Pile)

Max Activating Token Index: 58

Text #17

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.4709. Min Act: -0.0000

Data Index: 1713884 (The Pile)

Max Activating Token Index: 922

Text #18

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.5764. Min Act: -0.0000

Data Index: 64529 (The Pile)

Max Activating Token Index: 379

Text #19

Max Range: 5.1993. Min Range: -5.1993

Max Act: 3.4589. Min Act: -0.0000

Data Index: 976707 (The Pile)

Max Activating Token Index: 637