INDEX

Explanations

This neuron detects structured document headings or metadata field labels (e.g. “Course Description:,” “Prerequisites,” “Specification,” “LOCATION,” etc.).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disminución

0.44

ग्रियों

0.41

存在する

0.41

हेब

0.40

 obliv

0.40

<unused15>

0.39

 caminos

0.39

 prêts

0.39

 rapproche

0.38

 निवडणुकी

0.38

POSITIVE LOGITS

1.32

：

1.28

:}

1.04

1.02

ঃ

0.96

:**

0.95

:-

0.93

::

0.91

:</

0.91

:*

0.91

Activations Density 0.289%