INDEX

Explanations

programming identifiers and syntax

The neuron fires on capitalized code identifiers that name data‐type or API constructs (e.g. class/enum/struct names ending in words like “Property,” “Val,” “Request,” “Validator,” “Args,” or “Holder”).

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

↵↵

-1.33

are

-1.13

-1.12

 started

-1.09

 while

-1.08

for

-1.08

 from

-1.07

 ended

-1.05

-1.02

POSITIVE LOGITS

vement

1.32

というわけで

1.28

Deel

1.25

lblNewLabel

1.23

 renou

1.23

鋮

1.23

Är

1.22

Objekt

1.22

 апель

1.22

Lø

1.21

Activations Density 0.017%