INDEX

Explanations

backlog, proposal, decision, submission

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

 Paulo

-0.08

antal

-0.08

startIndex

-0.08

 Birch

-0.08

 practice

-0.08

Ø¹Ø§Ø¯

-0.08

 disqualified

-0.08

 stumbling

-0.08

ihar

-0.08

IFn

-0.08

POSITIVE LOGITS

 asset

0.13

 parameter

0.12

 talent

0.11

parameter

0.10

 product

0.10

 disease

0.10

acher

0.10

 consequence

0.10

 resource

0.09

 Parameter

0.09

Activations Density 0.113%