INDEX

Explanations

in terms of

The neuron is keyed to high‐level discourse connectives and sentence‐initiating function words (e.g. “However,” “To,” “Taking,” “Have,” “Excessive,” etc.), i.e. transition markers that structure or emphasize new clauses.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

一些

0.70

ちょっと

0.64

了很多

0.62

了一些

0.61

и

0.59

्ह

0.57

大概

0.57

Electric

0.57

 Exptl

0.57

ulfanyl

0.56

POSITIVE LOGITS

ാം

0.63

 convened

0.55

 sessão

0.54

 leaped

0.53

goers

0.52

 respectivas

0.51

 proclaims

0.51

 amid

0.50

 disagreed

0.50

 gegründ

0.50

Activations Density 0.002%