INDEX

Explanations

phrases indicating cause-and-effect relationships or results in various contexts

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

consequences of something

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

ChildScrollView

-0.64

inthians

-0.52

⟭

-0.51

 refroid

-0.51

igurumi

-0.51

 Technician

-0.49

Lithuan

-0.49

 tiguan

-0.48

AntiForgeryToken

-0.47

 féd

-0.46

POSITIVE LOGITS

 resulting

0.58

 CURIAM

0.55

 infolge

0.54

 Folge

0.54

 consequence

0.50

âce

0.50

 result

0.49

ECAUSE

0.48

 образом

0.47

__(/*!

0.46

Activations Density 0.013%

phrases indicating cause-and-effect relationships or results in various contexts

consequences of something

No Comments

No Known Activations