INDEX

Explanations

past, present, future, emerging

The neuron fires strongly on the conjunction “and” when it’s linking two time-related terms (e.g. “current and future,” “past and present”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

각

-1.34

-1.31

 jeweiligen

-1.30

并非

-1.30

astify

-1.29

 dichos

-1.24

 лек

-1.24

檄

-1.23

angsaan

-1.23

-1.22

POSITIVE LOGITS

 which

1.59

our

1.39

渡し

1.27

mgp

1.24

jao

1.21

 kwietnia

1.20

 Rät

1.19

寿

1.18

erobic

1.17

伱

1.16

Activations Density 0.018%