INDEX

Explanations

translation and type definitions

The neuron detects numeric tokens—digits, version numbers, IDs, counts, and other standalone numbers.

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

atzen

-0.79

this

-0.77

 δὲ

-0.76

 اجرای

-0.75

of

-0.75

ignation

-0.75

 aún

-0.74

Seznam

-0.73

ือ

-0.73

淬

-0.73

POSITIVE LOGITS

 type

1.17

 types

1.13

 типа

1.04

 typu

0.99

type

0.97

 untranslated

0.93

%),

0.92

%',

0.91

类型

0.90

 тип

0.90

Activations Density 0.002%