INDEX

Explanations

entities and roles

The neuron fires on generic category nouns naming “things” or “items” (e.g. attraction, site, product, food, item, piece, agent, destination).

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

bufio

-1.31

 alér

-1.25

 defStyleAttr

-1.20

the

-1.19

umán

-1.16

 můžete

-1.12

 gdzie

-1.11

getContentPane

-1.11

 всегда

-1.09

thd

-1.09

POSITIVE LOGITS

 both

1.69

of

1.48

by

1.38

for

1.37

ほとんど

1.34

if

1.33

ichtigen

1.33

‼️

1.30

 because

1.29



1.29

Activations Density 0.202%