INDEX

Explanations

mathematical or numerical contexts

The neuron is detecting negative numeric values (tokens that are a minus sign directly attached to digits).

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

 vetorial

-1.13

 doubtless

-1.10

 dourada

-1.09

 piscinas

-1.09

 scarcely

-1.07

 criaturas

-1.06

ビリ

-1.03

 encantador

-1.01

していない

-1.00

 Landwir

-0.99

POSITIVE LOGITS

cies

1.05

it

1.02

⁉

1.01

ZA

1.00

lihood

0.99

 Universite

0.99

 edifici

0.97

0.96

ers

0.95

智

0.95

Activations Density 0.030%