INDEX

Explanations

street, university, road, park names

The neuron responds to multiword proper‐noun phrases—especially named entities like organization or institution names, street/avenue/terrace addresses, and other capitalized place or group names.

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

鬭

-1.96

 canaux

-1.88

 étan

-1.88

 italiens

-1.86

 tâches

-1.76

 évé

-1.74

 清水

-1.74

 critères

-1.72

 mété

-1.72

 fonctionnalités

-1.72

POSITIVE LOGITS

to

2.55

 about

1.93

.'

1.84

$};

1.77

࿚

1.74

era

1.73

et

1.71

."

1.62

については

1.61

Activations Density 0.123%