INDEX

Explanations

phrases indicating growth or improvement in various contexts

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

increased measures

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

 подряд

-0.39

printList

-0.37

nerie

-0.36

asgi

-0.36

 nebo

-0.34

ItemClick

-0.34

 Révolution

-0.34

 defecto

-0.34

cupertino

-0.34

 benda

-0.34

POSITIVE LOGITS

posedge

0.76

脚注の使い方

0.71

$_['

0.67

 heightened

0.66

Increased

0.66

__(/*!

0.65

ngdoc

0.65

 increased

0.65

JspWriter

0.65

 ujednoznacz

0.64

Activations Density 0.033%

phrases indicating growth or improvement in various contexts

increased measures

No Comments

No Known Activations