INDEX

Explanations

only

This neuron fires on words (especially “only”) that signal an exclusivity or limitation in policy, pricing, or licensing statements.

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

 zarówno

0.94

ग्विजय

0.84



0.83

both

0.82

imilar

0.80

첩

0.78



0.77

daq

0.76

KAR

0.72

POSITIVE LOGITS

 лишь

1.91

 esclusivamente

1.22

 yalnızca

1.10

 apenas

1.09

 ओनली

1.08

 Hanya

1.06

僅

1.06

 ainult

1.04

しか

1.03

のみ

1.02

Activations Density 0.260%