INDEX

Explanations

implementation of

The neuron strongly activates on occurrences of the term “implementation” (often in the phrase “implementation of”) in policy or directive contexts.

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

[]:

-1.57

 необходимость

-1.42

</th>

-1.38

ambassade

-1.38

 nedeniyle

-1.35

 jiwa

-1.34

\\

-1.30

 інших

-1.25

 использованием

-1.23

"]="

-1.23

POSITIVE LOGITS

啣

1.98

 antena

1.52

 augusti

1.48

ୈ

1.39

רושלים

1.38

 主要

1.35

1.34

褙

1.33

塒

1.30

“)

1.30

Activations Density 0.038%