INDEX

Explanations

forward and backward direction

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ABCDEFGHIJKLMNOP

-0.11

evin

-0.11

 -âĢĲ

-0.10

aille

-0.10

enti

-0.09

ec

-0.09

izu

-0.09

oke

-0.09

utter

-0.09

kle

-0.09

POSITIVE LOGITS

-thinking

0.21

-looking

0.20

backward

0.17

ly

0.17

/back

0.17

-facing

0.15

-forward

0.15

ness

0.15

slash

0.14

-back

0.13

Activations Density 0.027%