INDEX

Explanations

lists and conjunctions

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

.linkedin

-0.12

/Linux

-0.12

_lhs

-0.12

.setLayoutManager

-0.12

lhs

-0.11

_LINUX

-0.11

 lasers

-0.11

 lavender

-0.11

 LinkedIn

-0.11

 lakes

-0.10

POSITIVE LOGITS

.trailingAnchor

0.16

unload

0.12

Unload

0.12

unl

0.11

 unload

0.10

 Dinner

0.10

 dinner

0.10

ListItem

0.10

_unlock

0.10

AllArgsConstructor

0.09

Activations Density 0.173%