INDEX

Explanations

already + state/possession

The neuron flags very high–frequency tokens—common words (including frequent verbs like “sitting” or “encrypting” and common adverbs like “already”) that recur across diverse contexts.

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

elijke

1.11

anath

1.01

}?

0.91

amam

0.89

wegian

0.88

Stick

0.88

ಮಾಣ

0.87

sion

0.86

aile

0.85

acles

0.85

POSITIVE LOGITS

 давно

1.61

 existente

1.38

既に

1.37

 bereits

1.26

Already

1.26

existing

1.23

すでに

1.20

 Existing

1.19

 existing

1.15

 esistono

1.14

Activations Density 0.069%