INDEX

Explanations

code identifiers and values

The neuron fires on identifier tokens being introduced or aliased—e.g. field names in SQL “AS” clauses, JSON property keys, type names, variable or import identifiers—i.e. on tokens that name or label a declaration or alias.

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

ue

-1.48

 dando

-1.43

 magnific

-1.41

 Dinas

-1.34

⤒

-1.31

andering

-1.30

othes

-1.30

 debacle

-1.25

ck

-1.23

 QUICK

-1.22

POSITIVE LOGITS

 EVERY

1.49

all

1.43

 overtly

1.41

 utterly

1.39

had

1.38

on

1.33

 went

1.30

 various

1.27

 meticulous

1.27

 every

1.25

Activations Density 0.011%