INDEX

Explanations

points to

The neuron fires on the words “point” or “points” when used in the phrase “point(s) to,” i.e. marks instances of that evidential/indicative construction.

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

-1.54

 mentre

-1.53

 Dicapai

-1.48

”

-1.44

 nachdem

-1.44

 '/';

-1.43

 hout

-1.42

those

-1.42

 tena

-1.38

while

-1.34

POSITIVE LOGITS

řízen

1.79

 potrebné

1.66

 ktorí

1.64

褡

1.64

 cierra

1.64

 knihy

1.59

 lanza

1.52

 obrá

1.51

 refugi

1.49

 váš

1.48

Activations Density 0.014%