INDEX

Explanations

height and dimensions

The neuron fires on numeric literal tokens (e.g. lengths or percentages) in code, i.e. it detects number values.

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

 menyukai

-1.36

 Organisator

-1.20

Since

-1.13

Koordinaten

-1.12

 menyes

-1.12

Following

-1.09

ත්

-1.08

 methods

-1.05

Being

-1.03

Generally

-1.03

POSITIVE LOGITS

の中では

1.23

1.12

迫力

1.09

 dreadful

1.03

 height

1.02

悻

0.97

 gewisser

0.97

 talvez

0.96

驚き

0.96

 somewhat

0.96

Activations Density 0.014%