INDEX

Explanations

code termination symbols

This neuron appears to be inactive—it does not reliably fire on any particular token pattern.

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

-1.91

to

-1.88

if

-1.54

for

-1.50

of

-1.41

˗

-1.39

 that

-1.38

 because

-1.35

,”

-1.34

 than

-1.33

POSITIVE LOGITS

Zubereitung

1.92

🕗

1.80

Beskrivning

1.75

ホント

1.59

Choisissez

1.58

replyreport

1.57

あぁ

1.57

Contenu

1.55

THEY

1.54

Descriere

1.54

Activations Density 0.001%