INDEX

Explanations

JSON, objects, arrays

This neuron activates on programming‐related identifiers and keywords (e.g. library paths, class or interface names, JSON/Data terms).

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

 either

-1.03

if

-0.98

 Keuangan

-0.98

ča

-0.96

 when

-0.94

 Studium

-0.94

at

-0.94

申请

-0.93

THREADS

-0.91

MessageWindow

-0.90

POSITIVE LOGITS

 icônes

1.19

 LIABLE

1.15

 kwanza

1.12

 objectifs

1.07

 côtes

1.06

<bos>

1.03

 réserver

0.99

ouncement

0.98

archical

0.96

 sélectionner

0.96

Activations Density 0.038%