INDEX

Explanations

equations with variables

This neuron activates on numeric and variable tokens (digits, letters used as variables) in arithmetic or algebraic expressions.

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

を行い

-1.22

 that

-1.22

-1.13

 bundes

-1.09

のでしょう

-1.06

...

-1.06

--

-1.05

以此

-1.05

 here

-1.01

POSITIVE LOGITS

 kalori

1.37

Technologie

1.31

توضیحات

1.24



1.23

Tahun

1.20

 rağmen

1.18

عنوان

1.18

Spesifikasi

1.17

 fön

1.17

 digunakan

1.16

Activations Density 0.139%