INDEX

Explanations

context and background

The neuron detects when the text is introducing or providing contextual background information.

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

șit

-1.45

 solchen

-1.42

 solche

-1.34

 miró

-1.30

adget

-1.27

 obstáculos

-1.23

 satisfecho

-1.22

one

-1.21

yyb

-1.18

Carex

-1.17

POSITIVE LOGITS

why

1.39

 which

1.35

 context

1.33

styleUrls

1.30

of

1.23

ificance

1.22

</strong>

1.20

Приложение

1.16

[]);

1.16

maxAge

1.12

Activations Density 0.094%