INDEX

Explanations

in locations or contexts

This neuron responds to high‐frequency function words—common pronouns, articles, prepositions, and similar grammatical tokens.

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

 Only

-1.33

之内

-1.30

τες

-1.27

 केवल

-1.27

avao

-1.24

 foolishness

-1.23

nteral

-1.23

 этим

-1.23

her

-1.22

<td>

-1.20

POSITIVE LOGITS

 that

1.96

and

1.96

 wektor

1.62

 hornear

1.56

倶

1.55

thenburg

1.49

 which

1.48

 then

1.48

 voeten

1.48

givings

1.46

Activations Density 0.056%