INDEX

Explanations

conjunctions indicating contrast

The neuron responds to longer, domain‐specific technical terms (especially in science/medical contexts).

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

__(/*!

-1.17

reszcie

-1.16

ないです

-1.16

читайте

-1.14

 powodu

-1.13

 alltså

-1.13

لقد

-1.11

：……

-1.09

:"-

-1.07

による

-1.03

POSITIVE LOGITS

 while

10.38

 mentre

7.38

 whereas

7.19

while

7.19

 mientras

7.09

 tandis

6.47

 enquanto

6.16

 whilst

5.72

While

5.50

而

5.28

Activations Density 0.550%