INDEX

Explanations

occasional or odd occurrences

The neuron highlights relatively uncommon content words—i.e. words with lower overall frequency or higher “importance” (TF–IDF–style) in the document.

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

-1.43

 peal

-1.35

 potentialities

-1.30

撺

-1.28

PAID

-1.27

覩

-1.26

 ovation

-1.24

-1.23

焢

-1.23

 spesifik

-1.23

POSITIVE LOGITS

 occasionally

1.62

 qualche

1.48

Occasionally

1.45

emplares

1.39

if

1.39

 sometimes

1.37

 begun

1.37

1.34

 algunos

1.34

干し

1.34

Activations Density 0.036%