INDEX

Explanations

ER, Office, auction, Controllers

This neuron fires on relatively rare, content-heavy tokens (uncommon or technical words) rather than on common function words.

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

 anglès

0.84

ন

0.83

CALL

0.78

 カラー

0.78

 国産

0.77

鉗

0.77

 sculpt

0.76

 Allora

0.76

 Baill

0.75

TOOL

0.73

POSITIVE LOGITS

it

0.88

 gilt

0.81

ler

0.80

os

0.79

ly

0.78

ged

0.78

ously

0.78

 sonucu

0.78

oder

0.77

ger

0.76

Activations Density 0.000%