INDEX

Explanations

seeking understanding or clarification

The neuron fires most strongly on second-person pronouns and direct reader addresses (e.g. “you,” “your”).

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

 what

-1.77

除此

-1.41

</h3>

-1.38

 wraz

-1.26

his

-1.23

our

-1.23

 itself

-1.23

-1.21

the

-1.20

あら

-1.19

POSITIVE LOGITS

 být

1.43

though

1.39

seite

1.38

 voelt

1.35

 betreft

1.33

wanderung

1.33

 interessanti

1.32

が欲しい

1.32

 мамы

1.31

 becauſe

1.29

Activations Density 0.015%