INDEX

Explanations

in cells

The neuron strongly activates on sentences that introduce or define a topic—that is, on definition‐style or encyclopedic explanations of a concept.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

дың

0.71

ның

0.70

0.63

 only

0.63

 pouze

0.63

0.62

<unused2121>

0.61

0.59

៧

0.58

POSITIVE LOGITS

けれど

0.81

但这

0.79

 لكن

0.70

但不

0.67

だけど

0.66

Some

0.66

BUT

0.66

Popular

0.64

 Nhưng

0.64

但也

0.64

Activations Density 0.006%