INDEX

Explanations

technical nouns and contexts

This neuron fires on the main subject or label term in headings or titles—i.e. the key noun that names the thing being described (e.g. “surface,” “missile,” “autoimmune,” “Standard,” “version,” “Gene,” “router,” “school,” “page,” “building”).

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

-2.14

lictions

-2.02

 only

-1.97

</b>

-1.91

</th>

-1.84

^{\

-1.84

’

-1.84

-1.80

 just

-1.78

and

-1.77

POSITIVE LOGITS

 perder

1.77

 involucrados

1.68

 besonder

1.67

 recomendamos

1.65

1.59

unikasi

1.57

有机会

1.55

1.53

 idyllic

1.52

Activations Density 0.207%