INDEX

Explanations

The neuron strongly activates on C/C++ preprocessor directive lines—especially “#include” statements and the header‐file names (in <…> or “…”) they reference.

import "

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

Italian

-1.47

 krás

-1.40

 nový

-1.39

interesse

-1.34

zura

-1.33

瘊

-1.32

 pères

-1.31

pantalón

-1.28

 nuova

-1.27

ándo

-1.26

POSITIVE LOGITS

1.29

 facil

1.11

1.09

 би

1.08

 Hobbes

1.05

 Calor

1.04

 ariel

1.02

 кри

1.02

Ov

1.01

It

1.00

Activations Density 0.029%