INDEX

Explanations

indicating finality or sufficiency

This neuron detects filler‐style constructions, especially the phrase “suffice it to say” (and similar “enough/only … to …” intensifier patterns).

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

 papir

-1.53

 eier

-1.45

 kupa

-1.36

 specifik

-1.35

cik

-1.31

 paik

-1.30

minValue

-1.30

 baseman

-1.27

any

-1.24

 Viertel

-1.23

POSITIVE LOGITS

 will

1.58

by

1.48

not

1.42

まずは

1.35

now

1.31

 afspraak

1.29

 Präsentation

1.26

нення

1.25

1.23

it

1.23

Activations Density 0.024%