INDEX

Explanations

parts of sentences

The neuron fires most strongly on the main predicate or “action” word (the central verb or claim) in each excerpt’s key 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

 then

-1.97

at

-1.48

 solely

-1.41

 into

-1.38

 merely

-1.27

-1.25

-1.24

 simply

-1.22

加拿

-1.21

liegenden

-1.21

POSITIVE LOGITS

%.

1.46

eleven

1.38

%;

1.38

}$.

1.33

⸂

1.32

GIVEREF

1.30

 utilizados

1.30

tling

1.30

ürnberg

1.29

Activations Density 0.263%