INDEX

Explanations

detailed

np_max-act · gemini-2.0-flash

questions inquiring for assistance or advice.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

This neuron activates on tokens that are part of user instructions or directives (e.g., “write,” “summarize,” “template,” “detailed plan”) indicating it’s detecting when the user is giving an instruction to the assistant.

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

ρος

-0.06

َأ

-0.06

 supermarkets

-0.06

 сообщ

-0.06

 omez

-0.06

 umož

-0.06

ụp

-0.06

.skip

-0.06

POSITIVE LOGITS

ény

0.06

.drawRect

0.06

 Glob

0.06

 provision

0.06

.psi

0.06

 probl

0.06

ANT

0.06

structures

0.06

워크

0.06

vog

0.06

Activations Density 0.076%

detailed

questions inquiring for assistance or advice.

This neuron activates on tokens that are part of user instructions or directives (e.g., “write,” “summarize,” “template,” “detailed plan”) indicating it’s detecting when the user is giving an instruction to the assistant.

No Comments

No Known Activations

detailed

questions inquiring for assistance or advice.

This neuron activates on tokens that are part of user instructions or directives (e.g., “write,” “summarize,” “template,” “detailed plan”) indicating it’s detecting when the user is giving an instruction to the assistant.

No Comments

No Known Activations