INDEX

Explanations

quantifiable metrics and statistical data

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

 surprise

-0.06

afil

-0.06

bak

-0.06

HEEL

-0.06

Gos

-0.06

 Mand

-0.06

 Julius

-0.06

 Ø¬Ø²

-0.06

-Semit

-0.06

chsel

-0.06

POSITIVE LOGITS

range

0.14

 range

0.13

 Range

0.13

 RANGE

0.12

Range

0.12

-range

0.11

(range

0.10

_range

0.10

	range

0.10

 subrange

0.10

Activations Density 0.047%