INDEX

Explanations

A, Ax, A_x

This neuron detects standalone mathematical coefficient symbols (especially uppercase letters like A or B, often with subscripts or attached numeric values) in equation contexts.

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

BitConverter

-1.27

-1.17

-1.16

 что

-1.13

nei

-1.12

 Nguyen

-1.11

 unseren

-1.10

𝑖

-1.10

_));

-1.09

Rabbi

-1.09

POSITIVE LOGITS

 there

1.52

 both

1.47

and

1.47

 because

1.43

hasilkan

1.40

 reduzir

1.39

as

1.36

 Following

1.32

噚

1.30

it

1.29

Activations Density 0.076%