INDEX

Explanations

commas followed by specific words

This neuron fires on content words that appear as items in an explicit example list (e.g. after “such as,” “for example,” “or the like,” “direct and indirect,” etc.).

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

但却

-1.45

人と

-1.34

%。

-1.31

TGA

-1.31

leggio

-1.30

notated

-1.30

 isnt

-1.27

 förslag

-1.26

-1.25

 cites

-1.24

POSITIVE LOGITS

it

3.13

we

2.19

 there

1.87

you

1.80

 this

1.72

 these

1.63

 resulting

1.51

these

1.46

are

1.46

 each

1.45

Activations Density 0.073%