INDEX

Explanations

under the

The neuron detects prepositional location phrases introduced by “under” or “beneath,” especially ones referencing the sky or overhead scenery.

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

皷

-1.63

to

-1.57

訁

-1.45

itair

-1.40

 pueden

-1.37

如果

-1.35

ри

-1.34

 هستند

-1.34

either

-1.33

 lingerie

-1.32

POSITIVE LOGITS

 što

1.55

 dengan

1.44

we

1.42

𓐍

1.41

the

1.38

 afecto

1.35

IGNS

1.34

 anuncio

1.30

 avec

1.30

様です

1.29

Activations Density 0.020%