JSON keys like title, description, version

This neuron activates solely on the double-quote character (")—i.e. it detects the quotation mark token.

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

菇

-2.23

 kabát

-2.08

 türkiye

-1.94

 menyen

-1.93

 parfü

-1.92

движи

-1.92

឵

-1.91

 pistolet

-1.89

 drapeau

-1.88

 brillant

-1.87

POSITIVE LOGITS

</strong>

2.38

2.33

2.31

2.25

2.23

2.19

2.14

2.11

Activations Density 0.043%