INDEX

Explanations

verify facts and claims

The neuron flags words and phrases that assert or question certainty, evidence, or proof—e.g. “proof,” “evidence,” “true,” “doubt,” and related terms.

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

 solstice

0.93

'></

0.91

ization

0.89

 windings

0.86

 tassels

0.85

 interface

0.84

0.83

 Preferences

0.81

tolist

0.80

زة

0.78

POSITIVE LOGITS

 доказа

1.43

 disprove

1.41

 preuves

1.40

 debunk

1.38

 allegations

1.36

證據

1.33

 거짓

1.32

 verità

1.28

 corroborated

1.28

 veracity

1.27

Activations Density 1.172%