INDEX

Explanations

starts many sentences

The neuron fires on the discourse-marker “So” when it begins or frames a sentence.

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

-1.30

 хотя

-1.14

 także

-1.13

 też

-1.13

 نیز

-1.10

 επίσης

-1.09

endphp

-1.09

]!=

-1.07

kijk

-1.01

 الرغم

-1.00

POSITIVE LOGITS

 there

1.77

 what

1.77

you

1.70

 they

1.45

 basically

1.44

 this

1.30

how

1.28

 then

1.27

we

1.24

now

1.20

Activations Density 0.025%