INDEX

Explanations

sequence progression and conclusion

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

adal

-0.10

 starting

-0.09

earer

-0.09

starting

-0.09

Starting

-0.08

 START

-0.08

 start

-0.08

 outr

-0.08

 Starting

-0.08

POSITIVE LOGITS

 progresses

0.21

 progressed

0.18

 continues

0.18

 progressing

0.18

 continue

0.17

 ends

0.17

 progressively

0.17

 proceed

0.16

 Continue

0.16

 gradually

0.16

Activations Density 0.296%