INDEX

Explanations

expressions of surprise or realization about unexpected outcomes

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ifest

-0.08

anel

-0.07

uese

-0.07

abort

-0.07

 nÃ©t

-0.07

 hete

-0.07

exter

-0.06

SU

-0.06

olist

-0.06

terra

-0.06

POSITIVE LOGITS

 such

0.07

å¦ĤæŃ¤

0.06

ãĥĪãĥ«

0.06

hest

0.06

/simple

0.06

 yoksa

0.06

 otherwise

0.06

 Turns

0.06

eh

0.06

such

0.06

Activations Density 0.004%