INDEX

Explanations

mathematical expressions and sequences

This neuron responds to non‐letter symbols—especially mathematical operators and punctuation (e.g. “=”, “+”, “,”, parentheses, question marks, braces, etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

selben

-1.56

 Scritto

-1.55

 素描

-1.37

㈰

-1.35

 culte

-1.32

注意事

-1.30

殼

-1.30

 电影

-1.27

 inget

-1.27

Цент

-1.26

POSITIVE LOGITS

has

1.58

was

1.55

に関して

1.30

 otras

1.30

at

1.30

 which

1.29

 other

1.27

may

1.25

can

1.24

 otros

1.24

Activations Density 0.014%