INDEX

Explanations

linking options and alternatives

The neuron fires on discourse‐level markers—personal pronouns and adverbs that signal a narrative or contextual shift (e.g. “my,” “his,” “suddenly,” “already,” etc.).

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

正好

-1.25

頑張って

-1.22

">

-1.13

of

-1.11

 overall

-1.09

 Româ

-1.09

 meren

-1.07

 akku

-1.07

aja

-1.05

ヤバ

-1.05

POSITIVE LOGITS

 nejen

1.28

 färg

1.20

 или

1.20

 both

1.19

 både

1.16

 möjlighet

1.14

 arbete

1.13

 måte

1.11

 либо

1.08

闪过

1.08

Activations Density 0.053%