INDEX

Explanations

issue number

The neuron responds to numeric tokens—especially multi‐digit numbers used as metadata (e.g. volume, issue, page, DOI numbers).

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

he

-2.03

もありますが

-1.86

 ваше

-1.77

邸

-1.77

poche

-1.73

 then

-1.69

および

-1.66

şi

-1.66

midos

-1.63

 喜欢

-1.62

POSITIVE LOGITS

 torner

1.78

 komplette

1.70

Pener

1.63

 Issue

1.63

Pemer

1.57

我只

1.56

Issue

1.56

 dalamnya

1.55

何故

1.54

Activations Density 0.010%