INDEX

Circuit Sparsity

No connected neurons data available

Explanations

random.ran with d/dint

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 256 tokens each

Dataset (Dashboard)

neuronpedia/python-code-simplified

Embeds

IFrame

Link

Negative Logits

...

-0.07

isi

-0.04

liver

-0.04

sert

-0.03

endencies

-0.03

Di

-0.03

enden

-0.03

move

-0.03

(di

-0.03

POSITIVE LOGITS

and

0.10

dint

0.06

der

0.04

If

0.03

 represe

0.03

list

0.03

 representing

0.03

 represent

0.03

ded

0.03

 contain

0.02

Activations Density 46.004%

No Known Activations