INDEX

Explanations

programming data types and structures

The neuron activates strongly on programming code tokens that name or manipulate data‐structure identifiers (e.g. “str”, “cell”, “list”, “double”, “expression” and quoted string names).

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

had

-1.44

であり

-1.37

 लिए

-1.34

ي

-1.30

łach

-1.27

 intensely

-1.26

ädigung

-1.26

and

-1.26

 essentially

-1.23

 actually

-1.23

POSITIVE LOGITS

 anfitrión

1.66

everyone

1.61

 namun

1.55

罔

1.51

.”

1.48

 capitán

1.46

჻

1.46

shtml

1.45

.’

1.42

 héroe

1.42

Activations Density 0.021%