INDEX

Explanations

endo- prefix or related roots

This neuron detects medical and scientific prefixes like "endo-", "endor-", and "Edmund/Edmond" with high activation.

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

 Ursache

-0.94

 exiled

-0.89

truded

-0.86

Doing

-0.85

 trung

-0.84

Comple

-0.83

 Inhibition

-0.82

 Erwä

-0.81

 remainder

-0.81

 einge

-0.80

POSITIVE LOGITS

 barbara

0.93

 сын

0.86

 NORMAL

0.85

瀟

0.85

 diam

0.84

TYP

0.84

 vell

0.83

らしい

0.83

 nico

0.82

 kawaii

0.82

Activations Density 0.058%