INDEX

Explanations

actual versus potential

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Comparative

-0.10

 slee

-0.10

aset

-0.10

uve

-0.10

agra

-0.09

ripe

-0.09

olders

-0.09

.Context

-0.09

usters

-0.09

 rede

-0.09

POSITIVE LOGITS

 coincide

0.13

 Actual

0.12

 coinc

0.12

 actual

0.10

Actual

0.10

actual

0.09

 discrepancy

0.09

_actual

0.08

 versus

0.08

 separate

0.08

Activations Density 0.306%

No Known Activations