INDEX

Explanations

each each each

The neuron fires on domain‐specific technical terms and abbreviations (e.g. chemical/drug names and acronyms) in scientific writing.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

or

-1.34

it

-1.25

 også

-1.23

 this

-1.22

 expérimentés

-1.20

its

-1.15

 например

-1.09

 apó

-1.00

 even

-0.99

了一眼

-0.98

POSITIVE LOGITS

각

1.48

 each

1.41

どれも

1.30

each

1.30

それぞれの

1.29

いずれ

1.27

 каждом

1.27

各个

1.25

各

1.20

各

1.18

Activations Density 0.087%