INDEX

Explanations

by [classification method]

The neuron fires on mailing-list navigation headers (e.g. “Next … by … thread,” “Previous … by … Date,” “message,” etc.).

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

 postgres

0.82

 endian

0.79

ipv

0.78

アルミ

0.78

 socialists

0.78

 observatory

0.77

 enduro

0.77

 evaluator

0.77

 auditing

0.76

 htonl

0.75

POSITIVE LOGITS

<h6>

0.66

 Função

0.62

Means

0.62

till

0.61

until

0.60

Philip

0.59

Committee

0.58

 frequência

0.58

Frog

0.57

ani

0.57

Activations Density 0.002%