INDEX

Explanations

performed two or beginning phrases

This neuron fires strongly on promotional or announcement language, highlighting show titles, event and content releases, and calls‐to‐action typical of marketing copy.

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

if

-1.88

or

-1.77

 правительства

-1.67

 السياس

-1.65

it

-1.48

ma

-1.45

 otro

-1.45

It

-1.45

io

-1.37

POSITIVE LOGITS

 fuis

1.77

Vgl

1.77

 دخترانه

1.68

 fizeram

1.65

 jejich

1.63

 joão

1.61

 their

1.60

 vírus

1.59

盉

1.58

 comprador

1.56

Activations Density 0.080%