INDEX

Explanations

states and conditions

The neuron activates on common nouns in their plural form.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

and

-1.70

-1.66

oczes

-1.41

zies

-1.33

haped

-1.30

an

-1.27

そして

-1.27

in

-1.22

on

-1.20

utenants

-1.20

POSITIVE LOGITS

 where

2.02

 that

1.68

Elaboración

1.58

ところに

1.42

為に

1.38

 zimowe

1.34

 considérée

1.33

 hvor

1.30

where

1.27

 tył

1.27

Activations Density 0.056%