INDEX

Explanations

associated with

The neuron detects expressions of formal affiliation or association (e.g. “affiliated with,” “associated with”).

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

nowu

-1.48

 fermer

-1.46

PropTypes

-1.44

fielder

-1.44

 partiellement

-1.42

 recevrez

-1.40

 trochu

-1.39

鯧

-1.39

 нравится

-1.38

笤

-1.38

POSITIVE LOGITS

of

2.72

 this

1.63

 событий

1.40

</strong>

1.37

 suggests

1.30

by

1.29

的一个

1.28

ingeki

1.26

 seemed

1.25

Activations Density 0.051%