INDEX

Explanations

mathematical proofs and statements

This neuron fires on the presence of LaTeX‐style mathematical notation (inline formulas, variables, and symbols).

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

 specif

-2.52

 cordones

-2.30

⚣

-2.25

 llavero

-2.05

 Etats

-2.03

 !!!!

-2.03

ſſen

-2.03

坶

-1.99

 créer

-1.96

THERE

-1.96

POSITIVE LOGITS

 with

2.64

or

2.42

all

2.25

 शुर

2.02

for

2.02

 Perhaps

1.99

as

1.98

 even

1.97

 before

1.96

 Surprisingly

1.94

Activations Density 0.090%