INDEX

Explanations

names joined by and

The neuron detects journalist bylines—that is, the personal names (and related credit tags) listed as article authors or reporters.

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

 pájaro

-1.27

 сахара

-1.22

 toalla

-1.09

而

-1.05

íslu

-1.05

 montón

-1.03

 zapato

-1.00

 maceta

-0.99

 that

-0.98

ׇ

-0.97

POSITIVE LOGITS

and

1.78

᾿

1.31

niña

1.13

 reporters

1.07

icago

1.05

 additional

1.05

ət

1.05

 الملف

1.04

}}{{

1.03

 addi

1.01

Activations Density 0.006%