INDEX

Explanations

of

np_max-act · gemini-2.0-flash

The neuron strongly activates on content-heavy nouns and nominal phrases—especially abstract “-ion” nouns (restoration, effect), formal treaties and institutions (Treaty, Empire), and other weighty event or entity names.

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

 Pháp

-0.06

 budete

-0.06

spam

-0.06

 Podcast

-0.06

jó

-0.06

 supportive

-0.06

ugu

-0.06

œur

-0.05

 Film

-0.05

_RTC

-0.05

POSITIVE LOGITS

UF

0.08

ада

0.08

 Conn

0.07

Notes

0.07

лер

0.06

ander

0.06

 متن

0.06

.wp

0.06

cht

0.06

rv

0.06

Activations Density 0.190%

of

The neuron strongly activates on content-heavy nouns and nominal phrases—especially abstract “-ion” nouns (restoration, effect), formal treaties and institutions (Treaty, Empire), and other weighty event or entity names.

No Comments

No Known Activations

of

The neuron strongly activates on content-heavy nouns and nominal phrases—especially abstract “-ion” nouns (restoration, effect), formal treaties and institutions (Treaty, Empire), and other weighty event or entity names.

No Comments

No Known Activations