INDEX

Explanations

common phrases and expressions

The neuron detects words that mark quoted speech or dialogue attribution—verbs like “said,” “says,” “say,” and similar reporting verbs.

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

 podrá

-1.62

 borracha

-1.55

ølge

-1.43

-1.41

↵↵↵

-1.40

 ABBREVIATIONS

-1.38

 malé

-1.38

);

-1.30

 Fähig

-1.30

 avait

-1.28

POSITIVE LOGITS

 these

1.72

 when

1.70

 These

1.59

 इन

1.54

這些

1.53

 этими

1.52

你不要

1.52

这些

1.50

姩

1.48

 هؤلاء

1.43

Activations Density 0.270%