INDEX

Explanations

occurs

This neuron activates on Russian-language (Cyrillic) text segments.

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

 pravi

0.74

帶著

0.69

 чуть

0.67

それを

0.66

 البته

0.65

だけど

0.62

こ

0.61

putting

0.61

 خودش

0.61

 czyli

0.60

POSITIVE LOGITS

出现

1.12

 došlo

1.11

 exists

1.06

 kullanılır

1.05

 возникает

1.05

 reikia

1.05

 existe

1.02

 можна

1.02

 galima

1.01

 необхідно

1.01

Activations Density 0.054%