INDEX

Explanations

numbers containing 3

This neuron activates specifically on the token “23” (i.e. it detects occurrences of the number 23).

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

 like

-0.88

 already

-0.87

 contrary

-0.86

 following

-0.79

 hashMap

-0.79

amaño

-0.78

 determined

-0.76

ăsp

-0.75

 ADVERTISEMENT

-0.74

hatic

-0.72

POSITIVE LOGITS

了她

0.88

DSP

0.87

фов

0.86

 heureuse

0.85

cea

0.81

ORING

0.81

—¡

0.80

ή

0.79

tigste

0.76

 élégante

0.76

Activations Density 0.001%