INDEX

Explanations

actions or mistakes

This neuron detects legal appeal language describing a court’s error (e.g., “erred,” “trial court,” “denied his motion”).

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

%;

-1.34

にほんブログ村

-1.34

mender

-1.21

pejo

-1.13

 Vainqueur

-1.12

:]:

-1.12

.},

-1.11

our

-1.10

 traductions

-1.09

 sabido

-1.09

POSITIVE LOGITS

 potrebbero

1.24

 щодо

1.21

and

1.21

chaqueta

1.17

 cuando

1.17

↵↵↵

1.17

などの

1.16

 merece

1.16

</b>

1.16

Activations Density 0.056%