INDEX

Explanations

knowledge cutoff date

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

-0.10

reau

-0.10

 fond

-0.09

inv

-0.09

sur

-0.09

THR

-0.09

 defect

-0.09

 pads

-0.09

 Seamless

-0.08

nder

-0.08

POSITIVE LOGITS

cut

0.11

 cutting

0.11

Cut

0.11

 cutter

0.10

 Cutting

0.10

Cut

0.10

cut

0.10

CUT

0.09

cuts

0.09

LineNumber

0.09

Activations Density 0.001%