INDEX

Explanations

`.items` attribute or `len` function (negative activations)

list comprehension indexing

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 256 tokens each

Dataset (Dashboard)

neuronpedia/python-code-simplified

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

ud

-0.09

ad

-0.09

verage

-0.08

che

-0.07

_si

-0.07

_paths

-0.07

_map

-0.06

yp

-0.05

_r

-0.05

lo

-0.05

POSITIVE LOGITS

Õ

0.40

point

0.26

 point

0.26

 vertex

0.25

 cell

0.25

_cell

0.25

vertex

0.24

 coord

0.24

_vertex

0.24

tile

0.23

Activations Density 99.999%