INDEX

Explanations

concrete nouns and abstract concepts

The neuron fires on concrete object or product nouns—terms naming items, devices, or other tangible things.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 with

-1.65

for

-1.51

 into

-1.43

 where

-1.42

llllll

-1.27

 will

-1.23

eeeeee

-1.23

and

-1.23

 futebol

-1.22

дир

-1.18

POSITIVE LOGITS

オリン

1.59

ようですが

1.58

 témo

1.52

 motivadoras

1.41

一方で

1.37

 externas

1.37

 privadas

1.34

ことが多い

1.34

 détaillés

1.33

 consideramos

1.32

Activations Density 0.104%