INDEX

Explanations

we

The neuron strongly responds to small, common function words (like pronouns or auxiliary verbs) in non-English (e.g. Spanish “lo,” Macedonian “сме,” Slavic verb forms).

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

but

-1.41

ዬ

-1.37

 Ottobre

-1.34

 atau

-1.30

 Gennaio

-1.30

-1.28

——”

-1.27

！”

-1.27

 Giugno

-1.27

itação

-1.26

POSITIVE LOGITS

 reaffirmed

1.70

蚨

1.70

釭

1.59

皤

1.58

 entered

1.57

糶

1.57

言えば

1.53

})}

1.49

 realizó

1.47

Activations Density 0.005%