INDEX

Explanations

chemical compound fragments

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

ìį¨

-0.09

.:.

-0.09

 Bias

-0.09

 preempt

-0.09

abus

-0.08

buz

-0.08

oldt

-0.08

 Verde

-0.08

.scalablytyped

-0.08

$MESS

-0.08

POSITIVE LOGITS

âĢ²

0.09

jiang

0.08

ooks

0.08

 trespass

0.08

 pubkey

0.08

à¸¯

0.08

_{}

0.08

áº¡o

0.08

bis

0.08

Activations Density 0.045%