INDEX

Explanations

position, boundary, center

This neuron fires on semantically rich “content” words (nouns, verbs, adjectives) rather than on small function words.

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

time

0.86

 συνεχ

0.80

 पूरा

0.77

0.76

 Sikh

0.70

 پورا

0.70

 الب

0.69

 świę

0.69

 Bukh

0.68

}}

0.68

POSITIVE LOGITS

ق

0.87

plante

0.84

ли

0.83

 lessening

0.82

 immédi

0.80



0.80

لاص

0.78

mités

0.77

লাষ

0.75

ль

0.75

Activations Density 0.001%