INDEX

Explanations

HTML tags, code comments

The neuron fires on numeric tokens and quantitative data (e.g. counts, measurements, IDs, percentages).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

나가

0.66

ные

0.61

返回值

0.60

起身

0.59

вит

0.58

眼鏡

0.58

setHeader

0.56

ClInclude

0.56

loadImage

0.56

setNome

0.56

POSITIVE LOGITS

<h2>

0.83

<h3>

0.77

år

0.60

<h4>

0.60

 serta

0.59

 devient

0.58

 deviennent

0.57

لې

0.56

 그리고

0.56

<h1>

0.55

Activations Density 0.137%