INDEX

Explanations

best practices

The neuron activates on the initial element of hyphenated compound terms (e.g. “meta-”, “follow-”, “long-”, “high-”).

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

ılan

0.68

xym

0.65

eture

0.64

َت

0.59

'}}

0.59

iburg

0.59

った

0.58

었

0.57

 XMFLOAT

0.57

boBox

0.56

POSITIVE LOGITS

ו

0.62

 wszystkich

0.61

🤝

0.61

 cómo

0.60

ଓ

0.59

↵↵

0.56

🌐

0.56

 לש

0.55

…,

0.54

 និង

0.54

Activations Density 0.001%