INDEX

Explanations

="

np_max-act · gemini-2.0-flash

The neuron strongly responds to the word “navigation,” especially in phrases like “Skip navigation links” or the “Navigation” label in HTML nav bars.

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

 talents

-0.07

 shielding

-0.07

�

-0.07

.multipart

-0.07

VML

-0.06

Bib

-0.06

 pymysql

-0.06

uji

-0.06

vou

-0.06

śmy

-0.06

POSITIVE LOGITS

={}

0.06

.assign

0.06

(freq

0.06

 alış

0.06

Storage

0.06

、、

0.06

dz

0.06

 lender

0.06

forum

0.06

 prendre

0.06

Activations Density 0.002%

="

The neuron strongly responds to the word “navigation,” especially in phrases like “Skip navigation links” or the “Navigation” label in HTML nav bars.

No Comments

No Known Activations