INDEX

Explanations

first release or hit

The neuron fires on adjectives or adjective phrases that highlight significance, order, or “first‐of‐its‐kind” milestones (e.g. major, first, real, full-length, big-budget).

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

羮

-1.85

蟜

-1.79

鈇

-1.73

鵙

-1.71

₋

-1.70

 generell

-1.69

lllllll

-1.67

⬮

-1.65

躂

-1.64

mediately

-1.62

POSITIVE LOGITS

of

2.11

2.09

 recurring

1.59

in

1.56

he

1.54

—

1.54

 thing

1.41

 increase

1.38

 amount

1.38

 array

1.38

Activations Density 0.043%