INDEX

Explanations

stay

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_19/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.19.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 HttpServletResponse

-0.08

Getting

-0.07

TableCell

-0.07

Led

-0.07

 seins

-0.07

 Rough

-0.06

bem

-0.06

}"

-0.06

 evening

-0.06

ValueHandling

-0.06

POSITIVE LOGITS

 strlen

0.08

前提是

0.07

檔案

0.07

.sum

0.07

 reopen

0.07

全局

0.07

加州

0.07

钞

0.07

 obscured

0.07

isActive

0.07

Activations Density 0.019%

stay

No Comments

No Known Activations