INDEX

Explanations

proper nouns or names

The neuron strongly activates on subword tokens that end in “o” (e.g. ‑aro, ‑ano, ‑ago, ‑avo), i.e. parts of proper names or words whose final character is “o.”

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

 either

-0.84

クロ

-0.77

{{/

-0.75

MOUSE

-0.74

̩

-0.73

setLoading

-0.72

 Teich

-0.70

 クロ

-0.69

 puffy

-0.69

">(</

-0.69

POSITIVE LOGITS

也知道

0.86

 föres

0.75

 espanhol

0.74

ographia

0.73

 PRACT

0.71

 réalité

0.71

ודם

0.71

 tačiau

0.69

 présent

0.69

worldwide

0.68

Activations Density 0.054%