INDEX

Explanations

linking words

The neuron spotlights adverbial or discourse‐marker phrases (e.g. “however,” “after all,” “in contrast,” “accordingly”) that signal transitions or emphasis in the text.

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

and

-4.63

in

-3.69

to

-2.92

as

-2.89

 including

-2.77

for

-2.73

at

-2.69

not

-2.69

 with

-2.64

by

-2.64

POSITIVE LOGITS

 refroid

3.16

 véhic

3.02

 réve

2.98

 bascul

2.81

 perpé

2.80

2.78

 garantir

2.70

 étan

2.66

 élar

2.63

 maî

2.58

Activations Density 0.404%