INDEX

Explanations

specific terms related to identification and categorization in structured data or systems

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

coordinates, location, number, description

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

IUrlHelper

-0.67

 protoimpl

-0.58

 للاسماء

-0.57

ThroughAttribute

-0.57



-0.56

+#+

-0.56

endous

-0.56

DIPSETTING

-0.55

saraba

-0.54

 BoxFit

-0.54

POSITIVE LOGITS

 landmark

0.33

 landmarks

0.33

IntoConstraints

0.32

 moverse

0.29

 puedas

0.27

 comunicarse

0.27

 consigo

0.27

 доступ

0.27

 colère

0.27

 Taktlose

0.26

Activations Density 0.087%

specific terms related to identification and categorization in structured data or systems

coordinates, location, number, description

No Comments

No Known Activations