INDEX

Explanations

phrases related to publication dates or original content

oai_token-act-pair · gpt-4o-mini Triggered by @bot

original publication date

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

GOTREF

-0.43

تقاوى

-0.42

Diweddarwch

-0.41

(~(

-0.41

➯

-0.39

 näch

-0.36

tinyos

-0.36

 mariage

-0.36

 linkovi

-0.36

 documentación

-0.36

POSITIVE LOGITS

 originalmente

0.53

bootstrapcdn

0.53

 published

0.53

Externé

0.52

 originally

0.52

 republished

0.51

 PUBLISHED

0.50

CloseOperation

0.50

原文

0.50

published

0.50

Activations Density 0.005%

phrases related to publication dates or original content

original publication date

No Comments

No Known Activations