INDEX

Explanations

mathematical and scientific terms

The neuron fires on segments containing technical mathematical notation—especially LaTeX‐style formulas, symbols, and terms (e.g. density, groups, “$…$,” angle brackets, numeric expressions).

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

 have

-1.67

of

-1.66

it

-1.63

was

-1.56

 would

-1.49

 They

-1.48

)$;

-1.44

枚目

-1.42

 таких

-1.42

↵↵

-1.42

POSITIVE LOGITS

2.00

 付き

1.52

あんまり

1.51

illots

1.45

凄く

1.45

珦

1.45

 chapéu

1.44

 siè

1.42

Припев

1.42

ほんと

1.41

Activations Density 0.110%