INDEX

Explanations

The neuron strongly responds to abstract, high‐level nouns denoting qualities or concepts (e.g. “majesty,” “precision,” “ability,” “significance,” “power”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

untuk

0.82

 लेकर

0.70

для

0.64

 উদ্দেশ্য

0.63

eniendo

0.63

 بعد

0.61

스를

0.61

ঘ্রই

0.60

려서

0.59

려고

0.59

POSITIVE LOGITS

of

1.54

 của

1.51

 ofthe

1.41

 của

1.34

 των

1.28

ของ

1.25

 ενός

1.22

 של

1.13

ຂອງ

1.09

ของการ

1.07

Activations Density 0.625%