INDEX

Explanations

engineering

The neuron fires strongly on tokens that denote specific numeric quantities or product/specification details (e.g. counts, measures, features).

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

lux

0.82

্যোগ

0.82

্তা

0.79

 chiếc

0.77

ккей

0.77

 rematch

0.76

 junit

0.75

ить

0.75

 mensagem

0.75

 لاکھ

0.74

POSITIVE LOGITS

ट

0.72

🤮

0.69

)}}

0.65

)**

0.64

 ajuste

0.63

!:

0.61

Hace

0.61

**.

0.60

 Даже

0.59

ઠી

0.59

Activations Density 0.000%