INDEX

Explanations

subjects of clauses

The neuron is activated by uncommon, content‐rich nouns and proper nouns (i.e. low‐frequency topical words).

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

and

-2.17

 также

-1.52

 también

-1.35

 également

-1.31

it

-1.27

 cercanas

-1.25

应用程序

-1.22

 yaklaşık

-1.20

{},

-1.17

為に

-1.16

POSITIVE LOGITS

が

1.45

でしょうね

1.41

 must

1.23

…？

1.20

Итак

1.15

いは

1.15

餉

1.13

ですかね

1.13

子は

1.13

だろうか

1.13

Activations Density 0.021%