INDEX

Explanations

numbers and units

The neuron is tuned to spot numeric tokens—quantities, measurements, and other numbers (including decimals and percentages)—in the text.

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

 waarvan

-1.24

⏟

-1.24

URERS

-1.22

 That

-1.21

一部分

-1.20

 apoi

-1.20

 less

-1.19

 THAT

-1.18

-1.17

了一些

-1.16

POSITIVE LOGITS

(>

1.49

 في

1.25

 even

1.20

 krä

1.17

 uusi

1.16

 даже

1.15

 любого

1.10

やっと

1.08

 nytt

1.07

 händ

1.07

Activations Density 0.111%