INDEX

Explanations

expressions of help or gratitude

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

followed this, wanted to

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

 komór

-0.35

+:+

-0.33

serviceWorker

-0.30

 cheese

-0.30

enderror

-0.30

Subsystem

-0.30

%?

-0.29

»?

-0.29

 Италијани

-0.28

 lojas

-0.28

POSITIVE LOGITS

 nahilalakip

0.69

 متعلقه

0.59

 lenker

0.53

 kasarigan

0.50

دانشنامهٔ

0.50

AndEndTag

0.49

IsContent

0.48

OGND

0.48

verifyException

0.48

のですね

0.47

Activations Density 0.024%

expressions of help or gratitude

followed this, wanted to

No Comments

No Known Activations