INDEX

Explanations

, question marks, and lists

The neuron detects names and branded entities (proper nouns and product/library names).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

-1.51

if

-1.36

 consists

-1.31

 from

-1.29

has

-1.25

-1.24

 tore

-1.20

 them

-1.19

si

-1.16

 consisting

-1.15

POSITIVE LOGITS

؛

1.45

LECTIONS

1.42

WERS

1.42

gä

1.39

 peringatan

1.37

 есть

1.34

kve

1.32

 этого

1.31

 современных

1.31

 этом

1.30

Activations Density 0.378%