INDEX

Explanations

end of clause or sentence

The neuron is specialized in spotting named organizations and institutional entities (e.g. company names, agencies, research centers, acronyms).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 different

-1.91

new

-1.75

 start

-1.48

 here

-1.43

 или

-1.42

 time

-1.41

 changes

-1.39

 including

-1.38

or

-1.37

 across

-1.36

POSITIVE LOGITS

 många

1.52

ら

1.45

водом

1.43

 kemer

1.43

 senin

1.41

 tirs

1.39

IALLY

1.33

ToMany

1.33

 المهم

1.32

めっちゃ

1.30

Activations Density 0.348%