INDEX

Explanations

HTML headings

The neuron detects section headings in HTML (text enclosed in <h1>–<h5> header tags).

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

 vår

-2.66

 restoran

-2.42

鰱

-2.39

the

-2.38

 annet

-2.30

iamo

-2.28

 vårt

-2.25

随即

-2.25

the

-2.23

其妙

-2.23

POSITIVE LOGITS

</h3>

2.53

 แต่

2.53

Of

2.34

 There

2.28

But

2.25

 According

2.22

 same

2.17

 otros

2.11

 very

2.08

大幅

2.05

Activations Density 0.002%