INDEX

Explanations

various objects or concepts

The neuron zeroes in on the main object or data structure being acted upon (the direct‐object term in questions and answers).

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

in

-2.47

at

-1.55

 have

-1.42

 each

-1.38

琑

-1.36

it

-1.34

 under

-1.31

𖥧

-1.27

 lät

-1.27

su

-1.23

POSITIVE LOGITS

 récemment

1.69

 عندما

1.43

神

1.39

 όταν

1.34

 когда

1.33

 lorsqu

1.32

 précédemment

1.30

 ktoré

1.30

 quando

1.28

 sekä

1.27

Activations Density 0.041%