INDEX

Explanations

opinion

np_max-act · gemini-2.0-flash

This neuron detects the judicial phrase marking the court’s official opinion (e.g. “the opinion of the court”).

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

 Marriott

-0.07

/issues

-0.07

 dames

-0.06

ゞ

-0.06

 slipping

-0.06

ior

-0.06

 Fruit

-0.06

loy

-0.06

(city

-0.06

Planning

-0.06

POSITIVE LOGITS

スマ

0.06

esát

0.06

ンティ

0.06

 ／:

0.06

 üçüncü

0.06

 závod

0.06

 kultur

0.06

습니다

0.06

.isOpen

0.06

ैं।

0.06

Activations Density 0.001%

opinion

This neuron detects the judicial phrase marking the court’s official opinion (e.g. “the opinion of the court”).

No Comments

No Known Activations