INDEX

Explanations

various studies indicate

This neuron fires on methodological or expository cues in technical writing—words and phrases that introduce procedures, examples, or study descriptions (e.g. “provided,” “various,” “for example,” “using”).

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

 aceasta

-1.28

 financiación

-1.27

 sostiene

-1.24

 attravers

-1.23

τός

-1.22

 giustizia

-1.22

完成です

-1.21

 solch

-1.17

 transportar

-1.17

 mostrare

-1.16

POSITIVE LOGITS

„

1.48

”

1.11

 mercredi

1.10

1.08

СКИ

1.06

“

1.05

SUB

1.05

 vendredi

1.04

ditionally

1.03

 jPanel

1.02

Activations Density 0.168%