INDEX

Explanations

attracting/attraction

np_max-act · gemini-2.0-flash

The neuron fires on words conveying the idea of drawing in or attracting (e.g. “attirait,” “attracted,” “drawn”).

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

 cigarettes

-0.07

urlencode

-0.07

 kıy

-0.06

 useCallback

-0.06

}')↵↵

-0.06

 Outlet

-0.06

 bgColor

-0.06

ást

-0.06

ेख

-0.06

üp

-0.06

POSITIVE LOGITS

 attracted

0.06

 český

0.06

Manufacturer

0.06

/code

0.06

 गलत

0.06

อเร

0.06

ulling

0.06

 joined

0.06

 resulted

0.06

 belongs

0.06

Activations Density 0.030%

attracting/attraction

The neuron fires on words conveying the idea of drawing in or attracting (e.g. “attirait,” “attracted,” “drawn”).

No Comments

No Known Activations