INDEX

Explanations

instances of the word "holding" and its variations, indicating a focus on physical grasp or attachment

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

olate

-0.07

ioni

-0.06

ucas

-0.06

uset

-0.06

rames

-0.06

ghi

-0.06

hn

-0.06

ihn

-0.06

POSITIVE LOGITS

alo

0.10

 onto

0.09

onto

0.09

-held

0.09

 hold

0.08

held

0.08

 Holding

0.08

 Holds

0.07

Ont

0.07

 HOLD

0.07

Activations Density 0.013%