INDEX

Explanations

numbers 1 and 9

The neuron strongly activates on numeric tokens (digits, scores, dates, statistics) regardless of context.

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

getText

-0.81

 bardziej

-0.78

holz

-0.75

xic

-0.73

 siete

-0.71

NEP

-0.69

 Nier

-0.68

 overall

-0.68

ði

-0.68

 Aufla

-0.68

POSITIVE LOGITS

 Ninety

0.93

 centenary

0.90

一百

0.81

mengen

0.81

extendable

0.74

 dagar

0.74

Unfinished

0.74

 ninety

0.74

 centen

0.73

 quen

0.72

Activations Density 0.023%