INDEX

Explanations

news stories

This neuron identifies meta-text labels in news articles—tokens like “story,” “edition,” and “headline” that mark article sections or links.

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

dorff

-1.37

ком

-1.37

VALUE

-1.36

+}$

-1.33

=\""

-1.31

GERS

-1.29

();

-1.28

-1.23

idase

-1.23

}$

-1.23

POSITIVE LOGITS

jsi

1.59

 about

1.55

うち

1.46

ופת

1.45

 всіх

1.43

 bezüglich

1.43

卽

1.42

 possono

1.39

 relacionados

1.38

变了

1.38

Activations Density 0.028%