INDEX

Explanations

foreign terms and grammatical connectors

The neuron selectively fires on software‐engineering and UI jargon—technical terms like “popup,” “inputs,” “refactoring,” “engagements,” “package,” etc.

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

-1.24

lze

-1.18

それにしても

-1.16

тию

-1.13

 группу

-1.12

板块

-1.12

 وهي

-1.10

-1.09

 there

-1.09

on

-1.09

POSITIVE LOGITS

apkan

1.15

Verweise

1.14

евна

1.14

 overal

1.13

Hao

1.13

illermo

1.11

 désir

1.10

 Volatile

1.09

camis

1.08

iroz

1.08

Activations Density 0.316%