INDEX

Explanations

the occurrence of the letter 'a' and its variants in different contexts

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

cta

-0.07

ct

-0.07

ward

-0.07

ir

-0.07

vig

-0.06

 Trick

-0.06

vier

-0.06

wise

-0.06

mann

-0.06

POSITIVE LOGITS

 also

0.08

also

0.07

 ÑĤÐ°ÐºÐ¶Ðµ

0.07

ãĤĪãģ³

0.07

SystemService

0.07

 takÅ¼e

0.07

isque

0.07

ÙĪÛĮÙĩ

0.07

 activeClassName

0.07

czy

0.07

Activations Density 0.004%