INDEX

Explanations

words followed by common descriptors

The neuron fires on the final word (or token) at the end of a sentence or chunk.

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

 should

-1.50

»:

-1.41

 robusto

-1.38

֥

-1.31

“:

-1.31

빠

-1.30

 stadig

-1.27

Имя

-1.27

 perumahan

-1.26

IDADES

-1.25

POSITIVE LOGITS

ッと

1.34

1.31

1.24

了出来

1.24

The

1.16

 regreso

1.13

нная

1.13

二是

1.13

 spécialisée

1.11

Until

1.10

Activations Density 0.128%