INDEX

Explanations

during time frames

This neuron fires on temporal markers—especially the word “during” (and similar time-indicating prepositions).

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

 afectada

-1.64

 that

-1.56

 Gefang

-1.52

して下さい

-1.49

 Erfol

-1.47

immen

-1.45

 Entsche

-1.44

而成

-1.44

ﾎ

-1.41

 Begleitung

-1.38

POSITIVE LOGITS

ors

1.54

Muchas

1.52

 richten

1.50

Tiene

1.45

茕

1.38

下的

1.37

 ‘‘

1.35

«

1.34

 decken

1.34

)':

1.33

Activations Density 0.091%