INDEX

Explanations

strong positive evaluations

The neuron flags evaluative or critical‐review language—i.e. words and phrases that express opinion or judgment about a work.

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

 then

-1.34

but

-1.31

 interesting

-1.30

 desagrad

-1.27

 wtedy

-1.26

 trochu

-1.23

 zawsze

-1.20

 beliebt

-1.20

AfterViewInit

-1.20

 разных

-1.19

POSITIVE LOGITS

 kado

1.56

kuh

1.48

 krim

1.41

 brille

1.38

 this

1.36

 vider

1.36

 konserv

1.35

 sorriso

1.34

 pedas

1.34

 boden

1.34

Activations Density 0.034%