INDEX

Explanations

things referring to objects or concepts

The neuron activates on the generic noun “thing” (and its plural “things”) when it appears in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 abbiano

-1.70

 after

-1.65

 avevano

-1.47

->___

-1.41

 manajemen

-1.41

 fallu

-1.39

ꠥ

-1.38

 uomini

-1.37

for

-1.34

InputBorder

-1.32

POSITIVE LOGITS

なども

1.62

などを

1.49

</h4>

1.48

 Infór

1.45

」、「

1.43

脸上的

1.39

 entire

1.38

etc

1.38

色々と

1.36

柔ら

1.35

Activations Density 0.027%