INDEX

Explanations

argues or states

The neuron is highly sensitive to named entities—especially people’s names, titles, and organizations (e.g. “Sgt Hindley,” “Transport Police,” speaker attributions).

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

 your

0.45

0.44

any

0.43

0.42

0.41

my

0.41

all

0.40

0.39

(=

0.39

POSITIVE LOGITS

 తెలిపారు

0.80

 বলেন

0.78

 말했다

0.77

 చెప్పారు

0.73

 said

0.72

 says

0.70

 જણાવ્યું

0.70

 mengatakan

0.69

said

0.68

 señaló

0.68

Activations Density 0.461%