INDEX

Explanations

consistency and contradiction

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

CodeGen

-0.08

unta

-0.08

ä¸įå½Ĵ

-0.07

æ§Ľ

-0.07

 strm

-0.07

gue

-0.07

åįĬè¾¹

-0.07

ufa

-0.07

utton

-0.07

åĤ¬åĮĸ

-0.07

POSITIVE LOGITS

 contradictions

0.12

 consistency

0.11

 contradictory

0.10

 inconsistencies

0.10

 inconsistency

0.10

 contradiction

0.10

 inconsistent

0.10

 coherence

0.10

istency

0.09

 Widers

0.09

Activations Density 0.020%