INDEX

Explanations

social, professional, or conceptual labels

The neuron fires strongly on copyright/legal disclaimer phrases (e.g. “All Rights Reserved”).

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



0.51

 infus

0.50

 परिश

0.49

केट

0.49

infra

0.49

kē

0.48

ayin

0.48

 आजम

0.47

ofy

0.47

ᴄ

0.47

POSITIVE LOGITS

 porém

0.57

 Porém

0.57

مج

0.56

൦

0.56

 jedoch

0.55

يد

0.55

espół

0.55

 père

0.54

 —,

0.54

؟

0.54

Activations Density 0.061%