INDEX

Explanations

official reports and inquiries

The neuron detects mentions of formal reports or reviews (e.g. commissioned inquiries and their titles).

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

pernicus

-1.28

 Aunque

-1.27

ान्त

-1.27

过来了

-1.27

糗

-1.26

-1.22

точный

-1.21

 Majesté

-1.20

Each

-1.19

 numérique

-1.19

POSITIVE LOGITS

 then

1.67

 that

1.67

text

1.38

 Haziran

1.38

 they

1.30

 econó

1.28

 just

1.26

 eventually

1.20

เรื่อง

1.20

end

1.18

Activations Density 0.057%