INDEX

Explanations

scholarships and awards

np_max-act · gemini-2.0-flash

The neuron is highly sensitive to structural “metadata” tokens in formal listings—section headings, numbered labels, years, and other numeric or colon-terminated labels delineating parts of the document.

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

р

-0.07

MP

-0.06

 dimensions

-0.06

М

-0.06

 house

-0.06

ię

-0.06

 taxed

-0.06

 cooking

-0.06

 rows

-0.06

 STATES

-0.06

POSITIVE LOGITS

 оформ

0.09

 prise

0.07

vyk

0.07

mex

0.06

 acompan

0.06

 आपक

0.06

čních

0.06

Pago

0.06

ead

0.06

.getLogin

0.06

Activations Density 0.008%

scholarships and awards

The neuron is highly sensitive to structural “metadata” tokens in formal listings—section headings, numbered labels, years, and other numeric or colon-terminated labels delineating parts of the document.

No Comments

No Known Activations