INDEX

Explanations

verbs followed by prepositions

The neuron spikes on short, high‐frequency “function” words—especially core verbs (like start, set, make, get, let, say, be) and pronouns (I, her, them).

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

もあって

-1.35

 fluide

-1.35

 唯美

-1.32

 手绘

-1.30

 イラストレーション

-1.29

 mauva

-1.28

 opérateur

-1.27

Dienst

-1.26

 bact

-1.26

 薬局

-1.25

POSITIVE LOGITS

 which

2.03

 from

1.92

to

1.89

 With

1.80

 with

1.62

 before

1.58

 Most

1.56

 Many

1.53

 that

1.52

 this

1.50

Activations Density 0.028%