INDEX

Explanations

beginnings of proper nouns

The neuron detects capitalized tokens that are the distinctive second (or later) element of multi‐word proper names—especially place names and personal 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

the

-2.64

-2.27

 both

-2.27

it

-2.20

but

-2.08

when

-1.91

 Since

-1.88

Many

-1.88

and

-1.88

POSITIVE LOGITS

 amie

2.13

̀i

1.96

 dezelve

1.93



1.90

 trajets

1.88

苼

1.88

袛

1.87

 ſame

1.87

 eſſ

1.83

 derselbe

1.82

Activations Density 0.194%