INDEX

Explanations

List separators and list items

The neuron strongly activates on words in their past‐tense or past‐participle form (the “-ed” suffix).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

0.38

0.35

dalam

0.34

Philippines

0.34

жды

0.33

تبر

0.33

tions

0.32

arıyla

0.32

दरअसल

0.31

during

0.31

POSITIVE LOGITS

 えっ

0.46

nó

0.42

มัน

0.42

 มัน

0.41

 그것

0.40

 かわ

0.40

it

0.39

wei

0.39

ᙱ

0.39

 impedir

0.39

Activations Density 0.244%