INDEX

Explanations

right side

The neuron detects mentions of small physical defects or damage—words like crack, chip, scar, bruise, stoma/scotoma, etc., often in a measured 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

erdere

-1.35

もしれない

-1.27

歟

-1.20

Ingredi

-1.20

も多く

-1.17

 レトロ

-1.16

Ꭳ

-1.16

 ダウン

-1.13

rechter

-1.09

Ồ

-1.09

POSITIVE LOGITS

one

1.64

1.18

on

1.13

 just

1.13

 unteren

1.09

 where

1.07

 several

1.07

 deportivas

1.02

 some

0.99

 יס

0.98

Activations Density 0.038%