INDEX

Explanations

the or eth

The neuron specializes in spotting proper names and other distinctive capitalized terms (e.g. uncommon names, titles, or brand/place names).

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

ete

0.69

etal

0.69

eta

0.64

eten

0.64

est

0.61

ev

0.61

 Metallurgy

0.61

ﺓ

0.61

enan

0.61

得

0.60

POSITIVE LOGITS

waite

1.14

lehem

0.93

some

0.84

olm

0.82

hhh

0.82

hhhh

0.81

hhhhhhhh

0.77

ttps

0.76

wick

0.75

étique

0.72

Activations Density 0.518%