INDEX

Explanations

dashed and solid lines

This neuron triggers on words that specify dashed or dotted line styles.

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

 accustomed

-1.50

VolleyError

-1.43

 muito

-1.38

 anúncio

-1.35

 информации

-1.35

 dados

-1.34

 cunning

-1.31

лиз

-1.31

 datos

-1.30

将她

-1.30

POSITIVE LOGITS

 esimerkiksi

1.48

niger

1.38

 $\&$

1.37

 formando

1.33

 fulf

1.33

1.29

 reccom

1.27

="

1.22

 duong

1.22

 Bring

1.20

Activations Density 0.011%