INDEX

Explanations

qualifier followed by concept

The neuron consistently lights up on the key technical terms (like “matter,” “energy,” “work,” “motion,” etc.) that appear as the main subjects in concise definition-style answers.

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

by

-1.27

at

-1.19

in

-1.17

んですが

-1.09

 same

-1.06

to

-1.00

 said

-0.98

 from

-0.94

But

-0.92

Objekt

-0.92

POSITIVE LOGITS

 bakteri

1.34

 cintur

1.26

 rokok

1.22

 plafon

1.21

 הַ

1.20

 caucho

1.17

 parfüm

1.17

 tarta

1.16

 tunik

1.16

 بیشتر

1.16

Activations Density 0.076%