INDEX

Explanations

expressions indicating complex character evaluations and moral dilemmas

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Negative Logits

reta

-0.08

ocus

-0.07

rire

-0.07

RET

-0.07

izu

-0.07

orget

-0.07

Disappear

-0.07

SingleNode

-0.07

ActionButton

-0.07

cret

-0.07

POSITIVE LOGITS

 underlying

0.10

 underneath

0.08

 deep

0.08

 ultimately

0.07

 vulnerable

0.07

 Bene

0.07

 human

0.07

 genuine

0.06

 deeper

0.06

 Geoffrey

0.06

Activations Density 0.025%

No Known Activations