INDEX

Explanations

asking if others have experienced this

The neuron detects community‐soliciting questions phrased like “Has anyone…?” or “Anyone else…?” asking if others have encountered or experienced something before.

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

 イラストレーション

-1.67

 barri

-1.57

墊

-1.41

镲

-1.40

 cauſe

-1.38

菒

-1.36

 Thessal

-1.36



-1.35

无论是

-1.34

 Teaser

-1.34

POSITIVE LOGITS

or

2.92

any

2.36

 this

2.13

 with

1.95

 from

1.94

 which

1.91

for

1.76

 longtime

1.68

 because

1.66

 gets

1.63

Activations Density 0.050%