INDEX

Explanations

if followed by pronoun

The neuron strongly activates on the conditional “if” (and related tokens like “then,” “you,” “is,” etc.), i.e. it detects conditional-clause triggers.

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

ma

-1.34

}$.

-1.34

 همچنین

-1.32

 belki

-1.32

jenigen

-1.32

 eachother

-1.30

пература

-1.30

푉

-1.28

ana

-1.27

",

-1.25

POSITIVE LOGITS

 this

1.47

gdy

1.43

it

1.40

präch

1.23

 soorten

1.22

罄

1.22

 samlet

1.21

section

1.21

線を

1.21

thenia

1.20

Activations Density 0.048%