INDEX

Explanations

looking at me or you

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

sic

-0.11

ewn

-0.09

 bran

-0.09

 vars

-0.08

AU

-0.08

 Aluminum

-0.08

Tow

-0.08

aza

-0.08

sle

-0.08

-summary

-0.08

POSITIVE LOGITS

 briefly

0.13

 directly

0.11

 smiling

0.11

 reaction

0.11

 through

0.11

through

0.10

Reaction

0.10

 Reaction

0.10

orting

0.10

 cand

0.10

Activations Density 0.028%