INDEX

Explanations

greeting new users

np_max-act · gemini-2.0-flash

The neuron activates on polite, conversational pleasantries—words and phrases used to express friendliness and cordiality (e.g. “pleasure,” “always,” “chat with you,” “thank you for asking,” etc.).

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

 Zend

-0.07

 Original

-0.06

>.</

-0.06

_SEC

-0.06

 swim

-0.06

 Palmer

-0.06

>"+↵

-0.06

_invoke

-0.06

 Room

-0.06

QtCore

-0.06

POSITIVE LOGITS

มากกว

0.07

exao

0.06

aneously

0.06

아서

0.06

per

0.06

 chví

0.06

Vũ

0.06

SO

0.06

ORMAT

0.06

azard

0.06

Activations Density 0.050%

greeting new users

The neuron activates on polite, conversational pleasantries—words and phrases used to express friendliness and cordiality (e.g. “pleasure,” “always,” “chat with you,” “thank you for asking,” etc.).

No Comments

No Known Activations

greeting new users

The neuron activates on polite, conversational pleasantries—words and phrases used to express friendliness and cordiality (e.g. “pleasure,” “always,” “chat with you,” “thank you for asking,” etc.).

No Comments

No Known Activations