INDEX

Explanations

applied to

This neuron detects terms related to applying or putting something into practice—words like “apply,” “applied,” “application,” “practical,” “relevance,” and “real-world.”

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

 hatten

-1.45

for

-1.43

了一会儿

-1.42

桩

-1.38

 estaban

-1.36

 время

-1.36

铅

-1.35

龃

-1.34

 levantó

-1.30

这

-1.29

POSITIVE LOGITS

</em>

1.83

1.59

1.43

 liefern

1.39

 kemudian

1.39

疡

1.38

1.34

'},

1.34

 занятия

1.32

“

1.31

Activations Density 0.041%