INDEX

Explanations

burg

np_max-act · gemini-2.0-flash

The neuron activates on tokens that are parts of German federal state or region names (e.g. “Me ck len burg,” “Sax ony,” “NR W,” “Mü ritz”), i.e. it’s looking for mentions of German states or regions.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 lance

-0.07

inicio

-0.06

WidgetItem

-0.06

Lif

-0.06

 nhân

-0.06

 savedInstanceState

-0.06

 Buckley

-0.06

 Robots

-0.06

orb

-0.06

.onError

-0.06

POSITIVE LOGITS

ク

0.08

,label

0.06

 Normal

0.06

 εργ

0.06

 germany

0.06

 Syracuse

0.06

há

0.06

GOT

0.06

ěž

0.06

िप

0.06

Activations Density 0.018%

burg

The neuron activates on tokens that are parts of German federal state or region names (e.g. “Me ck len burg,” “Sax ony,” “NR W,” “Mü ritz”), i.e. it’s looking for mentions of German states or regions.

No Comments

No Known Activations

burg

The neuron activates on tokens that are parts of German federal state or region names (e.g. “Me ck len burg,” “Sax ony,” “NR W,” “Mü ritz”), i.e. it’s looking for mentions of German states or regions.

No Comments

No Known Activations