INDEX

Explanations

explanation of why

The neuron detects causal and explanatory cue words—tokens like “explains,” “why,” “owing,” “result,” etc., that signal a cause‐and‐effect or explanatory relation.

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

DeclareMath

-1.11

ляли

-1.03

 シリーズ

-1.02

ρούν

-1.00

tly

-1.00

!!</

-0.99

:{}

-0.96

'/>

-0.95

 vuelva

-0.94

ഏ

-0.94

POSITIVE LOGITS

 some

1.17

so

1.10

xem

1.09

 algunas

1.04

 سیستم

0.99

我在

0.96

banyak

0.96

 overall

0.94

tiara

0.93

 někter

0.92

Activations Density 0.034%