INDEX

Explanations

JSON type definitions

The neuron fires on JSON/AST “type” identifiers—i.e. the names of object or node types (like “bundle,” “course-of-action,” “attribute,” “Document,” “newline,” etc.) in structured JSON or AST output.

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

-1.41

-1.37

-1.34

ot

-1.30

-1.29

するの

-1.29

-1.27

the

-1.21

ps

-1.21

POSITIVE LOGITS

 gestes

1.41

ängt

1.41

 only

1.41

敺

1.41

actrice

1.40

 itinéraires

1.36

 レモン

1.35

 neither

1.34

 leur

1.34

for

1.32

Activations Density 0.081%