INDEX

Explanations

chemical compounds and their properties

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

aned

-0.09

 bekan

-0.08

 Duffy

-0.08

ahn

-0.08

iw

-0.08

aber

-0.08

Uvs

-0.08

 Integral

-0.07

JsonValue

-0.07

 Swinger

-0.07

POSITIVE LOGITS

/-

0.10

âĢ³

0.10

_{}

0.10

lear

0.09

esson

0.09

zeigen

0.09

kening

0.09

 Wich

0.08

âĢ²

0.08

{:

0.08

Activations Density 0.079%