INDEX

Explanations

parts of a whole

The neuron activates strongly on the word “segment” (and its partial matches like “trechos” in Portuguese), i.e. it detects occurrences of “segment.”

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

-1.50

 vermek

-1.48

のかもしれません

-1.48

ouvrir

-1.45

世间

-1.44

efter

-1.42



-1.38

 описы

-1.38

زيل

-1.33

ſſel

-1.30

POSITIVE LOGITS

1.80

1.65

 причиной

1.50

."

1.48

なので

1.44

 любого

1.42

",

1.34

</em>

1.34

1.31

Activations Density 0.051%