INDEX

Explanations

as form, will remain, from outside

This neuron primarily responds to high-frequency function words—especially articles (“the”) and copular verbs (“is,” “are,” “be,” “comes”).

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

᾿

-1.74

ַּ

-1.67

året

-1.66

なども

-1.61

tissimo

-1.59

unfortunately

-1.59

 sowie

-1.58

 бампер

-1.57

ּוֹ

-1.56

POSITIVE LOGITS

 چیست

1.87

 only

1.87

 this

1.69

ׇ

1.58

 then

1.45

 coisas

1.45

ṓ

1.45

 fehlen

1.42

 рослин

1.41

 merely

1.40

Activations Density 0.042%