INDEX

Explanations

no reason or nothing

The neuron fires strongly on negative quantifiers—especially “no” and related words (e.g. “nothing,” “no apparent reason,” “no reward”) indicating absence or negation.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 کیلوم

-1.49

 involucrados

-1.47

晄

-1.30

ി

-1.29

全然

-1.22

 tirs

-1.22

 reworked

-1.21

 financières

-1.20

 times

-1.18

-1.16

POSITIVE LOGITS

 или

1.48

yorum

1.45

 vagy

1.40

 alebo

1.38

 логоти

1.35

1.34

となりました

1.33

 aDecoder

1.32

 chcete

1.32

quirer

1.32

Activations Density 0.011%