INDEX

Explanations

city/capitals

np_max-act · gemini-2.0-flash

The neuron activates on words indicating geographic or place‐type entities (e.g. “mount,” “city,” “capital”).

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

_k

-0.07

zent

-0.06

ismo

-0.06

—from

-0.06

�

-0.06

 zest

-0.06

rawer

-0.06

bubble

-0.06

—not

-0.06

attrib

-0.06

POSITIVE LOGITS

 신입

0.07

代码

0.07

][-

0.07

 قیمت

0.07

=Integer

0.06

 بیماری

0.06

WithIdentifier

0.06

 ApiException

0.06

")}↵

0.06

 oluştur

0.06

Activations Density 0.336%

city/capitals

The neuron activates on words indicating geographic or place‐type entities (e.g. “mount,” “city,” “capital”).

No Comments

No Known Activations