INDEX

Explanations

references to rewards and incentives in a transactional context

Tokens that serve functional rather than semantic purposes in text, often appearing as formatting elements, code syntax, punctuation, or structural markers at content boundaries.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

undle

-0.08

éļĨ

-0.07

APT

-0.07

.poly

-0.07

 nues

-0.07

ÏģÎ¿Î¹

-0.07

menin

-0.07

ickers

-0.07

ÄħÅ¼

-0.07

:async

-0.07

POSITIVE LOGITS

0.08

 tens

0.07

the

0.07

Âł

0.07

it

0.06

oreach

0.06

ãģĦãģĨ

0.06

an

0.06

but

0.05

Activations Density 0.111%