INDEX

Explanations

addresses, names, and locations

The neuron strongly activates on place names and other geographic or location entities (e.g. parks, roads, towns).

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

flats

-0.75

 admitir

-0.71

university

-0.70

 衣装

-0.70

organization

-0.69

 capteur

-0.68

 admitting

-0.67

越し

-0.67

 extensively

-0.67

 admit

-0.67

POSITIVE LOGITS

 пишу

0.73

 öde

0.73

BUCKET

0.71

 peripheral

0.71

 malzeme

0.70

Dut

0.70

彡

0.69

 Göttingen

0.69

личи

0.68

 пери

0.67

Activations Density 0.034%