INDEX

Explanations

programming code structure

The neuron selectively activates on numeric literal tokens (digits and numbers) in code.

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

 some

-1.25

 without

-1.23

 since

-1.18

any

-1.14

 even

-1.10

to

-1.06

 still

-1.05

isn

-1.01

by

-0.97

 before

-0.96

POSITIVE LOGITS

क्या

1.12

1.05

 подтвер

0.97

킁

0.93



0.93

TestBed

0.91

Santis

0.91

 risque

0.89

public

0.88

 helados

0.88

Activations Density 0.008%