INDEX

Explanations

states of being or opinions

The neuron fires strongly on the salient, content-bearing “key” words of a passage—that is, uncommon or thematically important terms that carry the most informational weight.

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

 ακό

-0.90

ποτε

-0.87

程式

-0.86

 đem

-0.83

гълү

-0.80

ventud

-0.77

лович

-0.77

 però

-0.77

 béné

-0.77

后来

-0.77

POSITIVE LOGITS

 regarding

1.61

这一点

1.45

 again

1.25

Regarding

1.17

 that

1.09

 here

1.09

这点

1.08

 concerning

1.06

这个问题

1.03

 this

1.00

Activations Density 0.049%