INDEX

Explanations

. After token

The neuron is looking for punctuation and connective cues (colons, commas, sentence-ending periods, contrastive conjunctions like “but,” parentheses, etc.) that mark boundaries or transitions in the text.

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

-1.40

to

-1.39

 мамы

-1.37

 recueil

-1.31

こういう

-1.30

these

-1.23

 each

-1.19

-1.16

を選ぶ

-1.16

いる

-1.13

POSITIVE LOGITS

gencias

1.76

 οποίο

1.71

tividade

1.54

"।

1.52

siębior

1.48

dicionado

1.48

 ہے۔

1.45

。「

1.45

وازم

1.41

sukienka

1.38

Activations Density 0.339%