INDEX

Explanations

er

The neuron strongly activates on blocks of text that read like a formal professional or biographical résumé—sections listing someone’s name, role, years of experience, and affiliations.

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

 flimsy

1.39

 jagged

1.25

 redness

1.20

 confusion

1.13

 ominous

1.13

 crumbs

1.12

 hasty

1.12

 changes

1.11

 não

1.11

 breakage

1.11

POSITIVE LOGITS

擔任

1.64

 Lecturer

1.52

 fondateur

1.48

担任

1.48

 работал

1.46

 متخصص

1.42

是一位

1.41

 Tätigkeit

1.41

 trabajó

1.40

 कार्यरत

1.38

Activations Density 4.103%