INDEX

Explanations

what you do

The neuron detects direct second‐person addresses (forms of “you,” “you’ll,” etc.) that speak to the reader.

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

of

-2.05

 even

-1.47

-1.40

 natürlich

-1.38

において

-1.34

 podrán

-1.32

 However

-1.30

 will

-1.29

 zelfs

-1.27

 tetapi

-1.23

POSITIVE LOGITS

it

1.63

 animados

1.57

 bonté

1.43

桫

1.40

 verksamhet

1.34

ícias

1.33

 quello

1.31

oved

1.29

 sagesse

1.28

quilo

1.27

Activations Density 0.189%