INDEX

Explanations

part and reasons

The neuron strongly activates on punctuation that marks sentence boundaries—especially periods (and to a lesser degree commas, question marks, 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

if

-0.94

咐

-0.90

イヤル

-0.88

Balancer

-0.85

chieht

-0.81

我国

-0.80

netje

-0.80

by

-0.79

ifty

-0.78

どんな

-0.77

POSITIVE LOGITS

 part

4.53

 reasons

3.25

part

3.19

Part

2.98

 Reasons

2.86

Reasons

2.66

 reason

2.59

 Part

2.58

 часть

2.53

 partly

2.52

Activations Density 0.094%