INDEX

Explanations

instruction and code

The neuron activates on decimal numbers (floating-point numeric tokens with a period).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

tions

0.88

grabbing

0.88

should

0.85

ными

0.85

year

0.83

type

0.82

trade

0.82

ную

0.81

ная

0.79

comparing

0.79

POSITIVE LOGITS

 blossomed

0.93

 Detalles

0.86

စိတ်အပိုင်း

0.85

 spiced

0.81

 pagine

0.80

箍

0.79

 capac

0.79

 revived

0.79

Paciente

0.78

 Poultry

0.78

Activations Density 0.000%