INDEX

Explanations

code directives

The neuron is detecting top‐of‐file or compiler/interpreter directives (e.g. include/import statements, `#pragma once`, `defined(…)` checks, PHP open tags, etc.).

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

 variedades

-1.38

any

-1.38

…

-1.31

 fábricas

-1.28

 große

-1.23

 first

-1.23

 большая

-1.22

 rantai

-1.20

 three

-1.19

claro

-1.19

POSITIVE LOGITS

1.77

1.76

1.57

 andra

1.54

場合には

1.48

1.43

1.38

1.37

1.34

Activations Density 0.009%