INDEX

Explanations

mathematical or logic clauses

This neuron detects words used in formal technical definitions or specifications to introduce options or examples (e.g. “may,” “various,” “include,” “such,” “types,” “groups,” “moieties,” “essentially,” “predetermined”).

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

 marea

-1.82

 perfe

-1.78



-1.77

 masaj

-1.73

 prato

-1.66

 ananas

-1.63

tev

-1.61

 pittores

-1.60

 paus

-1.60

 broek

-1.59

POSITIVE LOGITS

as

2.13

 like

1.75

at

1.75

 within

1.74

 before

1.67

 after

1.63

or

1.58

on

1.58

 While

1.57

 Before

1.57

Activations Density 0.157%