INDEX

Explanations

numbers and punctuation

This neuron fires on numeric reference tokens (volume, page, or section numbers) in scientific citations.

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

糴

-2.53

 Tentang

-2.06

träglich

-1.91

踯

-1.91

⢊

-1.90

 Terkait

-1.87

 veículos

-1.86

媗

-1.84

 Ubicación

-1.81

 Belajar

-1.80

POSITIVE LOGITS

1.89

AND

1.83

ol

1.83

in

1.81

！？

1.79

ar

1.78

ag

1.74

an

1.73

and

1.73

1.72

Activations Density 0.142%