INDEX

Explanations

concept names or descriptions

This neuron strongly activates on the distinctive content words that occur in titles or headings (e.g. key nouns and verbs central to a work’s name).

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

 that

-1.85

 until

-1.36

せずに

-1.26

 ketika

-1.21

 giudizio

-1.20

そう

-1.17

 appartamento

-1.13

 articolo

-1.11

 contratto

-1.10

 when

-1.09

POSITIVE LOGITS

 bästa

1.30

Ḥ

1.23

</tfoot>

1.21

('

1.16

 kvä

1.16

 största

1.15

šas

1.15

'\

1.14

 fyra

1.14

(=

1.12

Activations Density 0.822%