INDEX

Explanations

designs and types

The neuron detects technical specification terms—nouns naming design categories or quantitative parameters (e.g. “designs,” “configurations,” “levels,” “size”).

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

for

-1.37

 versie

-1.37

xxv

-1.33

блоки

-1.23

 kleuren

-1.21

 gepubliceerd

-1.20

ôs

-1.20

īgas

-1.18

 through

-1.16

 veiks

-1.16

POSITIVE LOGITS

 SUCH

1.30

 ujedno

1.29

 creer

1.28

鎂

1.26

賔

1.24

PENUTUP

1.22

rådet

1.20

鲲

1.19

跻

1.19

 если

1.18

Activations Density 0.112%