INDEX

Explanations

performing assessment

This neuron detects the term “assessment” (and its close variants like “assess” or “assessments”).

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

 Such

-1.55

to

-1.37

 Resultat

-1.29

CMAKE

-1.28

前的

-1.27

 Doesn

-1.27

mainloop

-1.27

 geomé

-1.24

 получила

-1.20

 Only

-1.20

POSITIVE LOGITS

out

1.53

if

1.50

 hermanas

1.48

籥

1.44

also

1.44

 другие

1.43

 there

1.41

 assessment

1.37

 OPINION

1.34

为了

1.33

Activations Density 0.080%