INDEX

Explanations

abstract qualities and characteristics

The neuron fires on abstract “property” nouns (e.g. format, type, style, substance, efficacy, functionality, power, etc.)—i.e. technical terms naming qualities or attributes.

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

in

-1.77

for

-1.65

𝒢

-1.22

 gobiernos

-1.20

 aldrig

-1.19

WAUKEE

-1.19

féle

-1.18

 Wherefore

-1.17

 courbe

-1.17

<td>

-1.16

POSITIVE LOGITS

its

2.02

and

1.95

those

1.48

</h4>

1.42

,:]

1.36

⑮

1.26

時間に

1.26

our

1.23

 zeitung

1.23

italien

1.20

Activations Density 0.069%