INDEX

Explanations

locations

The neuron is sensitive to prepositions like “on,” “off,” and “at” when they introduce specific location or address details.

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

 véritables

-1.87

readFileSync

-1.77

ープン

-1.77

勠

-1.70

 hän

-1.51

labili

-1.49

蹕

-1.48

ၜ

-1.48

 شیپور

-1.47

 sedih

-1.47

POSITIVE LOGITS

↵

1.54

to

1.53

‘

1.51

1.48

but

1.45

all

1.37

1.34

1.31

折り

1.30

 time

1.29

Activations Density 0.035%