INDEX

Explanations

HTML form elements

This neuron identifies programming‐related tokens—like code identifiers (e.g. Method, DateTime, start, label), punctuation (dots and hyphens), and class/group names—essentially marking bits of source code.

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

在那里

-1.34

 $^{\

-1.33

 xícara

-1.26

ﻒ

-1.25

natured

-1.25

 bolj

-1.24

 distritos

-1.23

 expresar

-1.23

 destinos

-1.21

 sejumlah

-1.20

POSITIVE LOGITS

1.66

櫚

1.60

”

1.37

name

1.37

 succulent

1.35

ޙ

1.34

 perfecting

1.34

1.33

…”

1.31

.“

1.30

Activations Density 0.014%