INDEX

Explanations

won races

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

vur

-0.96

 gruppen

-0.81

 følgende

-0.81

 criterio

-0.81

»>

-0.81

conductors

-0.79

 maquillage

-0.77

formations

-0.77

эффици

-0.77

無し

-0.77

POSITIVE LOGITS

 winner

1.42

win

1.16

 wins

1.16

 Winner

1.13

Winner

1.01

win

1.00

winner

0.99

WIN

0.98

placed

0.92

winning

0.92

Activations Density 0.004%

No Known Activations