INDEX

Explanations

tolerate, tolerance, bear

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

ext

-0.10

 cough

-0.10

ARAM

-0.09

::::::

-0.09

stub

-0.09

Disposable

-0.09

usty

-0.09

 Stub

-0.09

Bas

-0.08

POSITIVE LOGITS

 toler

0.51

 tolerance

0.46

tol

0.45

 bear

0.41

å¿į

0.39

 tolerated

0.35

bear

0.35

 Bear

0.35

 tolerant

0.32

 tolerate

0.32

Activations Density 0.098%