INDEX

Explanations

to [place name]

The neuron fires on verbs describing physical relocation or travel (e.g., “moved,” “travelled,” “headed,” “landed”).

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

 многих

-2.02

 mourir

-1.91



-1.83

 Dinas

-1.78

 usadas

-1.77

 wszystkich

-1.76

 refroid

-1.71

もあった

-1.69

 konkret

-1.67

 insanely

-1.67

POSITIVE LOGITS

an

2.28

by

2.27

 their

2.23

to

2.16

one

2.09

2.03

 with

1.96

for

1.82

 Since

1.66

 what

1.66

Activations Density 0.056%