INDEX

Explanations

seen or found followed by prepositions

The neuron activates on past-tense/participial verbs, especially “found” and “seen.”

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

for

-1.67

 significativos

-1.34

 ยาว

-1.32

三个

-1.30

six

-1.30

 three

-1.30

 teknis

-1.30

三人

-1.29

IRMED

-1.29

 drie

-1.28

POSITIVE LOGITS

1.51

↵↵

1.44

 gober

1.35

 unangemess

1.31

 любые

1.24

coledì

1.23

 Auffassung

1.23

 Absicht

1.20

if

1.20

 liegenden

1.20

Activations Density 0.079%