INDEX

Explanations

headings and code structures

The neuron is detecting HTML heading tags (e.g. `<h2>`, `<h3>`, `<h4>`, 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

”;

-2.70

“

-2.66

-2.39

-2.33

-2.27

ОВО

-2.23

腼

-2.22

 vår

-2.19

淡淡的

-2.19

POSITIVE LOGITS

最重要

2.67

not

2.59

to

2.55

now

2.52

潮流

2.42

harapkan

2.41

 allegedly

2.36

 admittedly

2.27

 generan

2.27

前任

2.23

Activations Density 0.000%