INDEX

Explanations

multilingual words

The neuron activates on words evoking warfare and combat, especially references to weapons and battle (e.g. “combat,” “rifle,” “steel,” “horseman”).

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

 turmoil

0.95

ੀਆਂ

0.93



0.93

 divisions

0.89

тные

0.88

oczy

0.87

resolvers

0.86

aways

0.85

ations

0.84

}^{+},

0.84

POSITIVE LOGITS

ną

1.25

ná

1.13

nod

1.13

 soyez

1.05

rd

1.05

 सेंटीमीटर

1.02

ge

1.02

nilai

1.01

che

1.00

َ

0.97

Activations Density 24.213%