INDEX

Explanations

This neuron is primarily firing on explicit numeric tokens (i.e. standalone digits or multi‐digit numbers).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Construct

0.72

 Disputes

0.71

THING

0.69

 Disorders

0.69

firstChild

0.66

 Jumping

0.66

 Molding

0.66

eqa

0.64

 superClass

0.64

subarray

0.64

POSITIVE LOGITS

 prenez

0.73

 améli

0.71

 alegre

0.71

 accompagner

0.70

 mesurer

0.68

퐈

0.68

 conforme

0.64

 conséquence

0.64

 méxico

0.63

 rêves

0.63

Activations Density 0.000%