INDEX

Explanations

this or that

The neuron fires on tokens that mark the ends of sentences or standalone lines, i.e. sentence‐ or block-final punctuation/words.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

*'

-1.27

 твое

-1.16

 твои

-1.07

kové

-1.02

 chini

-1.01

-.-

-0.96

ｌ

-0.96

 alho

-0.95

 beige

-0.94

 ngayon

-0.94

POSITIVE LOGITS

 this

1.66

はコチラ

1.32

droj

1.24

Кита

1.22

Удачи

1.18

 While

1.16

Выводы

1.14

 This

1.14

Миха

1.13

 that

1.12

Activations Density 0.016%