INDEX

Explanations

jaco taute /// inspector

This neuron fires on words that carry narrative content—place names, character and role names, and action verbs describing story events.

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

 maig

-1.40

Ār

-1.33

 kanton

-1.31

macht

-1.30

sich

-1.29

 artt

-1.27

 туры

-1.27

schon

-1.26

still

-1.25

 fabrik

-1.25

POSITIVE LOGITS

 државе

1.33

1.32

 this

1.31

鈑

1.30

 discussions

1.30

 like

1.28

<h6>

1.27

 think

1.27

 parecia

1.24

 beinhaltet

1.23

Activations Density 0.002%